Volume 14 Issue 2

AI-Powered Remote Work Fraud Detection System

Authors: Ms. Akanksha Patil, Aryan Gole, Harsh Pokharkar, Sanika Surve

Abstract: Due to the rise of remote work, ensuring employee accountability and performance has become increasingly challenging. This has also increased cases where employees misuse work hour using keyboard jigglers, auto-clickers, or embedded chips that simulate activity, allowing employees to appear active without actually working. This project presents an AI-driven system to detect such fraudulent behavior by analyzing keystroke dynamics, mouse movement patterns, and system activity logs. The goal is to build this system to detect such behavioural data including typing speed, key press duration, inter- key delay, mouse trajectory, and interaction timing to build a unique profile for each user. These behavioural patterns are then analyzed using machine learning algorithms to detect anomalies that indicate potential fraudulent activity. The System can keep the track of what people are actually doing on their devices in real time like how someone types or moves the mouse, the system can tell the work is being done by real person or just by using an automation tool. This project aims to improve transparency in remote job roles and promote honest work culture.

Analysis Of Heavy Metals And Health Risk Assessment Of Selected Energy Drinks In Nigeria

Authors: Hamza Abubakar Hamza, Ishiyaku Ibrahim Babayo, Ahmadu Muhammad Aliyu, Yusuf Mohammed Auwal, Abubakar Danjuma Bajoga, Hankouraou Seydou

Abstract: Energy drinks are widely consumed across Nigeria, particularly among youths, for their perceived ability to enhance physical performance and mental alertness. However, concerns have been raised about their safety, especially regarding heavy metal contamination, which may pose serious public health risks. Despite their increasing consumption, limited scientific data exist on the quality and safety of energy drinks sold locally. This study evaluated the concentration of heavy metals and assessed the associated health risks to consumers. Thirty (30) brands of energy drinks, including 23 liquid and 7 powdered samples, were randomly purchased from local markets. Heavy metal concentrations were measured using Atomic Absorption Spectroscopy (AAS) for liquid samples and X-ray Fluorescence (XRF) for powdered samples. Results were compared with World Health Organization (WHO) and European Food Safety Authority (EFSA) permissible limits. Cobalt (Co) concentrations ranged from 0.12 to 0.85 mg/L in liquid samples and 0.45 to 1.32 mg/g in powdered samples, with several powdered samples (EJ, KR) exceeding WHO limits. Chromium (Cr) concentrations were 0.08–0.67 mg/L in liquids and 0.35–1.10 mg/g in powders, while cadmium (Cd) ranged from 0.01–0.05 mg/L and 0.04–0.12 mg/g, respectively. Lead (Pb) levels reached 0.09 mg/L in liquids and 0.21 mg/g in powders, surpassing permissible limits in multiple brands, posing neurotoxic risks, particularly in children. Hazard Quotient (HQ) and Hazard Index (HI) values for Co, Cr, and Pb exceeded 1 in several samples, indicating potential non-carcinogenic risks. Carcinogenic risk (CR) values for Cr, Cd, Ni, and Co ranged from 1.2×10-4 to 5.8×10-4, exceeding the acceptable threshold of 1×10-4, suggesting a significant cancer risk, especially in children whose exposure per body weight was higher than adults. The findings indicate that certain energy drinks sold in Nigeria contain heavy metals at concentrations capable of causing both carcinogenic and non-carcinogenic health effects. These results underscore the urgent need for stricter regulatory oversight, routine monitoring of heavy metal content, improved manufacturing practices, and public education to mitigate health risks associated with excessive energy drink consumption.

DOI: https://doi.org/10.5281/zenodo.18938253

CyberSentinel: Fake Product Review Detection Using Machine Learning

Authors: V. Latha Sivasankari, Pratheep Kumar V, Preethika G, Pravin B

Abstract: Online marketplaces increasingly suffer from deceptive product reviews that manipulate customer perception and distort purchasing decisions. Traditional rule-based and manual moderation approaches struggle to detect sophisticated opinion spam, especially as review volumes grow exponentially across e-commerce platforms. The proposed system, Fake Product Review Detection Using Machine Learning, introduces an automated text analytics pipeline for identifying deceptive reviews using supervised learning techniques. The system processes raw review text through data preprocessing stages including tokenization, stop-word removal, normalization, and stemming, followed by feature extraction using TF-IDF vectorization. Multiple classification algorithms such as Logistic Regression, Naïve Bayes, and Support Vector Machine (SVM) are evaluated to determine optimal performance. A trained model is integrated into a Flask-based web application that enables real-time review classification as Fake or Genuine. The system architecture ensures seamless interaction between preprocessing, feature engineering, model inference, and user interface components. Performance evaluation conducted on a labeled dataset demonstrates an accuracy of 85%, with balanced precision and recall values, confirming reliable detection capability. The modular Python-based implementation ensures scalability, maintainability, and ease of deployment on standard computing environments. This approach enhances trustworthiness in online review ecosystems by providing an efficient, intelligent, and automated fake review detection solution.

DOI:

INFLUENCE OF Nano Silica On Mechanical Properties Of Concrete_733

Authors: Vivek D, Shri Hari Varshaa S, Nisha Gopika D

Abstract: The construction industry increasingly demands sustainable materials with enhanced performance characteristics to address environmental concerns and structural requirements. This experimental investigation evaluates the influence of nano silica as a partial cement replacement on the mechanical properties of concrete mixtures. The study employed a systematic approach where nano silica was substituted for Ordinary Portland Cement at replacement levels of 0%, 1%, 2%, and 3% by weight, with concrete specimens prepared using standard fine and coarse aggregates following established mixing procedures. Comprehensive mechanical testing was conducted to assess compressive strength at 7, 14, and 28 days, split tensile strength using cylindrical specimens, and flexural strength through four-point bending tests on beam specimens. The results demonstrated significant improvement in mechanical properties with optimal performance observed at specific replacement percentages, indicating enhanced strength development and structural performance compared to conventional concrete. The findings reveal that nano silica replacement not only reduces cement consumption for environmental benefits but also provides superior mechanical characteristics, establishing viable dosage ranges for practical construction applications and contributing to the development of sustainable, high-performance concrete formulations for modern infrastructure projects.

An Integrated Agentic AI Framework For Personalized Academic Assistance In Technical Education_819

Authors: Dr.M.Senthil Kumar, Mrs. S. Vimala, Darshwana R S, Dhayalini S S

Abstract: This paper introduces an integrated agentic framework leveraging generative artificial intelligence (AI) to provide comprehensive academic assistance in technical education. The system utilizes a multi-agent architecture powered by large language models (LLMs) and retrieval-augmented generation (RAG) to deliver personalized support including syllabus-aware tutoring, intelligent study planning, and real-time attendance analytics. By grounding generative responses in department-specific datasets (CSE, AI-DS, ECE), the framework mitigates hallucinations and ensures curriculum alignment. Experimental validation with 150 students across three engineering departments demonstrates an 87% satisfaction rate, a 65% reduction in information-seeking time, and a 42% improvement in study plan adherence. The results suggest that agentic workflows can significantly enhance student engagement and administrative efficiency in higher education environments.

DOI: https://doi.org/10.5281/zenodo.18952340

Recent Solar Events And High-Energy Cosmic Ray Particles

Authors: Rekha Agarwal, Rajesh Kumar Mishra, Divyansh Mishra

Abstract: Solar Cycle 25 (SC25), the most energetically active solar cycle in approximately two decades, has produced a remarkable sequence of solar events that have profoundly modulated the flux of high-energy cosmic ray (CR) particles near Earth. This paper presents a systematic, multi-instrument analysis of the effects of recent SC25 events — including five Ground-Level Enhancements (GLEs), extreme Forbush Decreases (FDs), and an Anisotropic Cosmic-Ray Enhancement (ACRE) — on high-energy galactic cosmic rays (GCRs) and solar energetic particles (SEPs) detected by the global neutron monitor (NM) network, the SEVAN multi-directional detector, and space-based platforms. We reconstruct the differential rigidity spectra of SEPs during GLE73–GLE77, characterise the rigidity-dependent amplitude of major FDs across stations spanning Rc = 0.01–13 GV, and demonstrate a fundamental spectral asymmetry between GLE enhancements (hard SEP component concentrated at R ~ 1–4 GV) and FD suppressions (selectively removing the softer GCR component below ~5 GV). GLE74 (11 May 2024), embedded within the largest FD of SC25 (−15.7% at Oulu NM, 10 GV rigidity) and a G5 geomagnetic storm, exhibited a moderately soft spectrum (γ ≈ 5.0–6.3) with a spectral rollover above ~2 GV and an unusually broad angular distribution indicative of pitch-angle scattering in the highly disturbed heliospheric field. GLE77 (11 November 2025), the strongest event in 19 years, displayed a harder spectrum (γ ≈ 4.5–5.0) and a characteristic double-peak structure attributable to distinct flare-impulsive and CME-shock-accelerated particle populations, confirmed by the first simultaneous gamma-ray detection at both Arctic and Antarctic polar stations. The ACRE event of 5 November 2023 — only the third known event of its kind and the first detected to 8 GV midrigidity — demonstrates that complex heliospheric magnetic configurations can focus and intensify high-energy GCRs from the anti-Sun direction at energies above 10 GeV. These findings collectively advance the physical understanding of CR acceleration, transport, and modulation in the inner heliosphere during a strong solar maximum and carry direct implications for radiation protection in aviation, human spaceflight, and space weather operations.

A Study On Peak Load Pricing And Sustainable Demand Management In Tourism

Authors: Ananya, Dr.M.D.Chinnu

Abstract: On one hand, tourism destinations around the world are dealing with excessive demand in peak seasons. The local infrastructure suffers a great deal as a result of environmental degradation. As a result, the visitors’ satisfaction has decreased. This study analyses the potential of peak load pricing, as an economic instrument, for sustainable demand management in tourism. Peak load pricing refers to the pricing of a higher price during the peak period. And a lower price during the off-peak period. To control the flow of tourists and not merely maximise revenue. Through the analysis of the interconnection between pricing strategies and sustainable tourism management, this article opines that peak load pricing can be effectively used to balance economic growth with environmental protection.

A Rural Telemedicine Network For Remote Healthcare Delivery Using AI-Assisted Consultation And Digital Prescription Systems

Authors: Dr.M. Senthilkumar, Mrs.A. Sangeetha, Hareini S, Kowsiga Shri P, Jayadharshini A

Abstract: This paper presents the design, architecture, and implementation of A Rural Telemedicine Network aimed at bridging healthcare access gaps in geographically isolated communities. The proposed system integrates Interactive Voice Response (IVR) telephony for feature phone users, smartphone-based video consultation, AI-assisted doctor information retrieval, digital prescription generation, and pharmacy verification through QR-code authentication. A microservices architecture hosted on cloud infrastructure ensures scalability and low-latency operation in bandwidth-constrained environments. The platform supports multi-role access encompassing patients, doctors, health workers, pharmacies, and administrators. Experimental deployment across three rural pilot sites demonstrated a 78% reduction in patient travel time, a 91% prescription verification accuracy, and over 85% user satisfaction among health workers. The results validate the feasibility of affordable, resilient telemedicine for underserved populations and underscore the transformative potential of combining AI, and mobile health technologies in primary healthcare delivery.

DOI:

Digital Transformation And Advanced Technologies In Supply Chain Management: Empirical Evidence From Omani E-Commerce Companies

Authors: Khoula Al Kaabi, Dr. Zeeshan Asim, Noor Al Yahyai, Fatema Al Isaee, Ali Saif Said Al Maqbali, Hyder Kamran

Abstract: The study seeks to investigate the impact of digital transformation and emerging technologies on the supply chains of Omani e-commerce businesses. E-commerce is an emerging sector in the country; however, the sector is facing a lot of challenges in supply chain logistics, infrastructure, and fluctuating demands. Technologies such as Artificial Intelligence, Internet of Things, Big Data, Blockchain, and Cloud Computing have the potential to impact supply chains. The study aims to explore the positivism paradigm, and the data will be collected through a survey from supply chain professionals along with secondary research that will be collected from peer-reviewed journals. 5 hypotheses have been developed for the study which talks about the rela-tionship of emerging technology with supply chain digital transformation. The results of the study say that the awareness level about the emerging technologies is high among the supply chain professionals, however the implementation level is low due to costs, skills, and resistance. Strategic recommendations are provided at the end of the research for e-commerce businesses, government, and supply chain professionals.

DOI: https://doi.org/10.5281/zenodo.18958697

Green Transportation And Economic Performance In Food Supply Chain: An Operational Perspective

Authors: Hajer Abdul Hameed Haji Al Zadjali, Dr. Zeeshan Asim, Nasser Said Nasser Al Bdul Salam, Mohammed Majid Ahmed Al Marzuqi, Dr. Devarajanayaka Muniyanayaka

Abstract: This paper will explore economic performance of supply chain and green transportation in transportation from an operations and strategic fit perspective. Contrary to traditional economic performance of supply chain and green transportation in transportation which addresses the cost of sustainability, this paper will address the operational aspect of sus-tainability in supply chain and green transportation in transportation which can be aligned with economic performance through supply chain strategic fit which accounts for the inherent uncertainty and demand-response nature of supply goods. Sustainability as an operation rather than a cost will be aligned with economic performance through supply chain strategic fit which addresses product requirements through an efficiency-responsiveness frontier.

DOI: https://doi.org/10.5281/zenodo.18958843

Supply Chain Management Practices And Supply Chain Performance: Empirical Evidence From The Retail Sector Of Oman

Authors: Hajer Abdul Hameed Haji Al Zadjali, Dr. Zeeshan Asim, Usama Hamed Ali Alnofli, Arshad Mehmood, Hyder Kamran

Abstract: The purpose of this study is to examine the effect of supply chain management practices on supply chain performance in the retail industry of Oman. In this study, six supply chain man-agement practices such as strategic supplier relationships, customer relationships, information sharing, information technology, employee training and internal operations have been taken as independent variables for supply chain performance. The quantitative research method has been used in this study. The primary data has been collected with the help of a structured question-naire from 71 retail industry professionals of Oman. The Cronbach’s alpha method has been used for reliability testing, and the hypotheses of the study have been tested through Pearson correlation method. The findings indicated that the five SCM practices, namely customer rela-tionships, information sharing, information technology, employee training, and internal opera-tions, have a statistically significant and positive impact (R² = .673; Adjusted R² = .581) on supply chain performance. Strategic supplier relationship practice was found to have a non-statistically significant impact, which is in contrast with most of the literature, and this is a poten-tial research gap for future studies. This study fills the literature gap in the relatively unexplored Omani retail sector and offers some guidelines for retailers who want to improve their operation-al performance, lower their costs, and increase their competitive advantage.

DOI: https://doi.org/10.5281/zenodo.18958990

Data Science: An Overview And Its Applications

Authors: Hari Murugan U, Muhammed Raashidh I, Dr.K.Brindha

Abstract: Data Science has emerged as one of the most influential disciplines of the 21st century due to the exponential growth of digital data generated every day. It integrates statistical analysis, computer science, machine learning, and domain expertise to extract meaningful insights from large and complex datasets. Organizations across industries—such as business, healthcare, finance, and social media—leverage data science to improve decision-making, optimize operations, and create personalized user experiences. With approximately 2.5 quintillion bytes of data produced daily, the need for effective data management, analysis, and interpretation has become essential. This paper explores the fundamental concepts of data science, including data collection, preprocessing, exploratory data analysis, machine learning, deep learning, and big data technologies. It also discusses commonly used tools and platforms such as Python, R, TensorFlow, and cloud-based machine learning services that support scalable data analysis and model deployment. Furthermore, the paper highlights diverse applications of data science across sectors such as healthcare, agriculture, e-commerce, and smart cities.

DOI: https://doi.org/10.5281/zenodo.18975795

Title: Navigating The Ethical Landscape Of Generative AI In The Pharmaceutical Sector: Embedding An Ethical Framework To Ensure Responsible Innovation

Authors: Harkish Sen

Abstract: Generative Artificial Intelligence (Gen AI), Pharmaceutical Industry, AI Ethics, European Union Artificial Intelligence Act (EU AI Act 2024), Drug Discovery, Clinical Development, Data Privacy, Transparency, Accountability, Equity, Responsible AI, Regulatory Compliance, Ethics by Design.Predictive Analytics, Data Visualization, Cloud Computing, Explainable AI, AutoML.

DOI: https://doi.org/10.5281/zenodo.18976621

Leveraging Geospatial Analytics And Machine Learning For Precision Business Expansion: A Micro-Market Framework

Authors: Mrs. Sangeetha Priya, Tarun G, Prasanth Pm

Abstract: The strategic expansion of business operations is a critical yet high-risk endeavor, often hampered by a reliance on macro-level market data that fails to capture localized nuances. This paper addresses this challenge by introducing a comprehensive framework for micro-market analytics, designed to facilitate data-driven, precision-targeted business expansion. The proposed model integrates multi-source data— including demographic, geospatial, transactional, and psychographic information—to segment large urban areas into distinct micro-markets. By applying machine learning algorithms, specifically a gradient boosting model, the framework generates a “Market Potential Score” to quantify the success probability for each granular location. The methodology is validated through a hypothetical case study of retail expansion in a Tier-2 Indian city, demonstrating its ability to identify high-potential, low-risk opportunities that traditional analysis would overlook. The framework culminates in a visualization dashboard, providing stakeholders with an intuitive tool for strategic decision-making. This approach significantly enhances the precision of expansion strategies, minimizes financial risk, and promotes sustainable business growth in competitive environments.

DOI: https://doi.org/10.5281/zenodo.18977153

Route Analysis And Traffic Jam Prediction Using Deep Learning Techniques: An Enhanced LSTM-GRU Comparative Study

Authors: Ohm Prakasanatham

Abstract: Traffic route analysis plays a critical role in daily urban mobility. This study proposes and evaluates deep learning models for predicting traffic jam status based on contextual weather and event data. Specifically, LSTM and GRU architectures are trained and tested on a dataset of weather conditions, air quality, road characteristics, and events. This study introduces a structured hyperparameter optimization framework and an interpretive analysis explaining observed performance gaps. Data preprocessing includes one-hot encoding, Min-Max normalization, and temporal sequence organization. Both models are evaluated using accuracy, precision, recall, F1-score, and ROC AUC. The LSTM model achieves 87.5% accuracy and 91.67% ROC AUC, significantly outperforming the GRU model (50% accuracy, 67% ROC AUC). The findings offer practical guidance for selecting appropriate recurrent architectures in traffic prediction systems. Traffic prediction, deep learning, LSTM, GRU, weather data, event data, traffic congestion, performance evaluation, hyperparameter optimization.

DOI: https://doi.org/10.5281/zenodo.20320705

<img

Safe-Sight

Authors: Parth Bhalekar, Aarya Darne, Sakshi Devadiga, Mrs. Akanksha Patil

Abstract: The absence of immediate visual data presents significant obstacles for individuals with visual impairments when navigating unfamiliar settings. Although they provide some degree of support, conventional assistive technologies, such as white canes, guide dogs, and basic smartphone applications, often exhibit limitations in terms of functionality, cost, or real- time adaptability. To address these limitations, this study proposes Safe-sight, a mobile application using artificial intelligence. The goal is to improve the safety and independence of people with visual impairments. The system uses text-to- speech technology to provide immediate audio feedback. It also uses optical character recognition to read text, and object detection, which is powered by deep learning models. Voice command is the simplest way to interact, allowing for use and easy control. The result highlight the system’s ability to improve mobility, situational awareness, and confidence for users with visual impairments. Safe-sight, therefore, provides a solution that’s both practical and scalable, while also being budget- friendly, merging accessibility with the latest AI developments.

AI Powered Legal Document Analyser

Authors: Prof. Tejashree pangare, Tanvi Patil, Pratham Surve

Abstract: The AI-Powered Legal Document Analyser is an Android application created to assist users in understanding complicated legal documents including contracts, agreements, and legal notices. Many legal documents contain complex terminology that is hard for non-expert professionals to interpret. This application enables users to upload or scan documents, after which Optical Character Recognition (OCR) retrieves the text and Natural Language Processing (NLP) techniques examine the content to detect key clauses, responsibilities, potential risks, and time limits. The application also includes a legal chatbot that responds to user queries, generates easy summaries of the document. Built using Java and XML for the Android interface and Firebase for safe data storage, the system maintains user privacy while offering fast and dependable legal insights. This project aims to make legal information easier to access, affordable, and easy to understand for individuals, students, and small businesses.

Ethanolic Leaf Extract Of Guava Leaves (Psidium Guajava) Against Staphylococcus Aureus And Escherichia Coli

Authors: Dr. Samson L. Mangin

Abstract: Guava (Psidium guajava) is a fruit-bearing tree known for its numerous health benefits. Its leaves are widely recognized for their medicinal properties and have long been used in traditional medicine. In emergency situations, guava leaf extracts have been commonly utilized as antibacterial agents and have been applied in the treatment of common wound infections. This study aimed to determine which concentration of guava leaf extract is most effective as an antimicrobial agent against Staphylococcus aureus and Escherichia coli, which are common causes of wound infections. The experiment began with the extraction of dried guava leaves at concentrations of 50%, 75%, and 95%. Mueller–Hinton agar was used to culture the two pathogenic bacteria along with the extracted guava leaf solutions. Each bacterium was exposed to the three different concentrations of guava leaf extract, while penicillin was used as the control. The cultures were then incubated at 37°C for 24 hours to observe the zones of inhibition. After 24 hours, the results showed that guava leaf extract produced measurable zones of inhibition, indicating antimicrobial activity. The 50% guava leaf extract showed the lowest zone of inhibition against Staphylococcus aureus, with an average of 15.67 mm. In contrast, the 95% guava leaf extract exhibited the highest zone of inhibition against Escherichia coli, with an average of 22.33 mm. Based on these findings, the study conclude that guava leaf extract is effective as an antimicrobial agent against Staphylococcus aureus and Escherichia coli.

DOI: https://doi.org/10.5281/zenodo.18996214

Smart Aerial Spraying System : An IoT-Integrated Quadcopter For Sustainable Farming

Authors: Mrs.K.G. Suhirdham, M. Ganesh, S. Gunasivan, P. Manoj

Abstract: The growing demand for sustainable agricultural practices calls for innovative technologies that enhance productivity while reducing environmental impact. This study introduces a Smart Aerial Spraying System, an IoT-enabled quadcopter developed to support precision farming operations. The system integrates unmanned aerial vehicle (UAV) technology with real-time sensing and cloud-based monitoring to deliver accurate and efficient pesticide and fertilizer application. Embedded sensors collect data on temperature, humidity, soil moisture, and crop conditions, enabling adaptive spray control based on field requirements. Through IoT connectivity, farmers can remotely supervise flight operations and spraying parameters using a mobile or web dashboard. GPS-guided navigation and automated route planning ensure uniform coverage and operational safety. The proposed solution reduces chemical overuse, lowers labor dependency, and minimizes human exposure to hazardous substances. Experimental results indicate enhanced spraying accuracy, better resource utilization, and improved efficiency compared to traditional manual methods, highlighting its potential as a reliable tool for sustainable and intelligent agriculture.

DOI: https://doi.org/10.5281/zenodo.18997708

Multi-Class Pneumonia Detection Using Machine Learning and Deep Learning Approach

Authors: Dr A. Rajaprabhu, Sameksha V R, Sanjanaa K, Sariga K

Abstract: Pneumonia is a life-threatening respiratory infection that affects millions of people worldwide, requiring accurate and early diagnosis to reduce mortality rates. This paper presents a multi-class pneumonia detection system using Machine Learning (ML) and Deep Learning (DL) approaches to classify chest X-ray images into Normal, Bacterial Pneumonia, and Viral Pneumonia categories. The proposed framework integrates image preprocessing techniques such as resizing, normalization, and data augmentation to enhance model robustness. Traditional ML classifiers, including Support Vector Machine (SVM) and Random Forest, are compared with advanced Convolutional Neural Network (CNN) architectures for performance evaluation. Feature extraction is performed using both handcrafted features and deep feature representations. The system is trained and validated on publicly available medical imaging datasets, and evaluation metrics such as accuracy, precision, recall, and F1-score are used to assess performance. Experimental results demonstrate that deep learning models outperform conventional machine learning algorithms in multi-class classification tasks, providing higher diagnostic accuracy and reliability. The proposed approach can assist radiologists in early pneumonia detection and improve clinical decision-making.

DOI: https://doi.org/10.5281/zenodo.18997856

Plant Disease Detection And Monitoring Using Artificial Neural Network

Authors: Anne Debora P, Narendran S, Dr. S. Sheeja

Abstract: Fungi have been identified as a major threat to crop production in the world. In this study, methods of improving the performance of plant disease detection and prediction using artificial neural network techniques are presented. The hyperspectral fungi dataset of 21 plant species were collected and trained using back propagation algorithms of an artificial neural network to improve the conventional hyperspectral sensor. The system was modelled using self-defining equations and universal modelling diagrams and then implemented in the neural network toolbox in Matlab. The system was tested validated and the result showed a fungi detection accuracy of 96.61% and the percentage increment was 19.53%.

DOI: https://doi.org/10.5281/zenodo.18998367

Microalgae-Based Biofertilizers: A Promising Alternative To Chemical Fertilizers

Authors: Shuchita Pandey, Poornima Devi, Parul Singh, Dr. Amita Pandey

Abstract: Microalgae have emerged as valuable assets in agriculture, serving as biofertilizers and soil conditioners. In tropical lowland rice cultivation, nitrogen-fixing cyanobacteria-based biofertilizers have proven effective, while eukaryotic, unicellular, green microalgae find application in temperate zones for soil conditioning, particularly in sprinkler-irrigated farmland for erosion control. The potential of microalgae technologies is substantial, yet their success varies due to challenges such as the absence of quality inoculates and a limited understanding of soil microbial ecology. These obstacles hinder the development of integrated management schemes crucial for maximizing the effective exploitation of microalgae in crop production. This review delves into the current state of microalgae applications in agriculture, addressing prospects and challenges across various domains, including crop production, protection, and natural resource management. It emphasizes the need for concerted screening and strain improvement programs, coupled with advancements in product formulation technologies, to overcome existing limitations. Furthermore, the review provides an overview of recent advances, novel technologies, their commercialization status, and outlines future directions. By shedding light on the potential and hurdles, this comprehensive analysis aims to guide efforts towards realizing the full spectrum of benefits that microalgae can offer in enhancing agricultural practices.

Awareness Of Post Office Savings Schemes Among The Public In Coimbatore: An Empirical Study On Knowledge Levels, Investment Behaviour, And Accessibility Barriers

Authors: Mohammed Anshaf.A, Dr. S.V. Harshini

Abstract: Walk into any post office in Coimbatore on a weekday morning and you will see the same thing: a queue of people waiting to pay electricity bills, pick up parcels, or send money home. Very few are there to ask about the Public Provident Fund or the Senior Citizen Savings Scheme. That is not because those products are bad — by almost any measure, they are excellent. It is because most people simply do not know they exist, or know only enough to be unsure where to start. This paper examines exactly that problem — the awareness gap around Post Office savings schemes — using survey data from 100 residents of Coimbatore district, collected between November 2024 and March 2025. We kept the analysis simple: structured questionnaire, simple percentage analysis, and four theoretical frameworks to make sense of the patterns. What we found was striking enough that we think it deserves a clear, direct write-up rather than one buried in statistical complexity. The headline number is this: 70% of respondents do not have full awareness of the Post Office savings schemes available to them. Word-of-mouth from family and friends is carrying most of the awareness load at 36%, while social media — which dominates how this city’s under-40 population discovers almost everything else — accounts for just 6% of scheme awareness. Safety and government backing motivate two-thirds of investment decisions. Service quality is where the experience lets people down most. And yet, 68% of those surveyed said they would recommend Post Office schemes to someone they know. There is genuine goodwill here. It is just not being channelled effectively. For anyone working in financial inclusion, public savings mobilisation, or India Post’s operations in cities like Coimbatore, these findings point to a clear and correctable problem. The post office has the trust, the products, and the physical infrastructure. What it is missing is communication that actually reaches people.

Awareness Of Post Office Savings Schemes Among The Public In Coimbatore: An Empirical Study On Knowledge Levels, Investment Behaviour, And Accessibility Barriers

Authors: Mohammed Anshaf.A, Dr. S.V. Harshini

Abstract: Walk into any post office in Coimbatore on a weekday morning and you will see the same thing: a queue of people waiting to pay electricity bills, pick up parcels, or send money home. Very few are there to ask about the Public Provident Fund or the Senior Citizen Savings Scheme. That is not because those products are bad — by almost any measure, they are excellent. It is because most people simply do not know they exist, or know only enough to be unsure where to start. This paper examines exactly that problem — the awareness gap around Post Office savings schemes — using survey data from 100 residents of Coimbatore district, collected between November 2024 and March 2025. We kept the analysis simple: structured questionnaire, simple percentage analysis, and four theoretical frameworks to make sense of the patterns. What we found was striking enough that we think it deserves a clear, direct write-up rather than one buried in statistical complexity. The headline number is this: 70% of respondents do not have full awareness of the Post Office savings schemes available to them. Word-of-mouth from family and friends is carrying most of the awareness load at 36%, while social media — which dominates how this city’s under-40 population discovers almost everything else — accounts for just 6% of scheme awareness. Safety and government backing motivate two-thirds of investment decisions. Service quality is where the experience lets people down most. And yet, 68% of those surveyed said they would recommend Post Office schemes to someone they know. There is genuine goodwill here. It is just not being channelled effectively. For anyone working in financial inclusion, public savings mobilisation, or India Post’s operations in cities like Coimbatore, these findings point to a clear and correctable problem. The post office has the trust, the products, and the physical infrastructure. What it is missing is communication that actually reaches people.

Awareness Of Post Office Savings Schemes Among The Public In Coimbatore: An Empirical Study On Knowledge Levels, Investment Behaviour, And Accessibility Barriers

Authors: Mohammed Anshaf.A, Dr. S.V. Harshini

Abstract: Walk into any post office in Coimbatore on a weekday morning and you will see the same thing: a queue of people waiting to pay electricity bills, pick up parcels, or send money home. Very few are there to ask about the Public Provident Fund or the Senior Citizen Savings Scheme. That is not because those products are bad — by almost any measure, they are excellent. It is because most people simply do not know they exist, or know only enough to be unsure where to start. This paper examines exactly that problem — the awareness gap around Post Office savings schemes — using survey data from 100 residents of Coimbatore district, collected between November 2024 and March 2025. We kept the analysis simple: structured questionnaire, simple percentage analysis, and four theoretical frameworks to make sense of the patterns. What we found was striking enough that we think it deserves a clear, direct write-up rather than one buried in statistical complexity. The headline number is this: 70% of respondents do not have full awareness of the Post Office savings schemes available to them. Word-of-mouth from family and friends is carrying most of the awareness load at 36%, while social media — which dominates how this city’s under-40 population discovers almost everything else — accounts for just 6% of scheme awareness. Safety and government backing motivate two-thirds of investment decisions. Service quality is where the experience lets people down most. And yet, 68% of those surveyed said they would recommend Post Office schemes to someone they know. There is genuine goodwill here. It is just not being channelled effectively. For anyone working in financial inclusion, public savings mobilisation, or India Post’s operations in cities like Coimbatore, these findings point to a clear and correctable problem. The post office has the trust, the products, and the physical infrastructure. What it is missing is communication that actually reaches people.

Awareness Of Post Office Savings Schemes Among The Public In Coimbatore: An Empirical Study On Knowledge Levels, Investment Behaviour, And Accessibility Barriers

Authors: Mohammed Anshaf.A, Dr. S.V. Harshini

Abstract: Walk into any post office in Coimbatore on a weekday morning and you will see the same thing: a queue of people waiting to pay electricity bills, pick up parcels, or send money home. Very few are there to ask about the Public Provident Fund or the Senior Citizen Savings Scheme. That is not because those products are bad — by almost any measure, they are excellent. It is because most people simply do not know they exist, or know only enough to be unsure where to start. This paper examines exactly that problem — the awareness gap around Post Office savings schemes — using survey data from 100 residents of Coimbatore district, collected between November 2024 and March 2025. We kept the analysis simple: structured questionnaire, simple percentage analysis, and four theoretical frameworks to make sense of the patterns. What we found was striking enough that we think it deserves a clear, direct write-up rather than one buried in statistical complexity. The headline number is this: 70% of respondents do not have full awareness of the Post Office savings schemes available to them. Word-of-mouth from family and friends is carrying most of the awareness load at 36%, while social media — which dominates how this city’s under-40 population discovers almost everything else — accounts for just 6% of scheme awareness. Safety and government backing motivate two-thirds of investment decisions. Service quality is where the experience lets people down most. And yet, 68% of those surveyed said they would recommend Post Office schemes to someone they know. There is genuine goodwill here. It is just not being channelled effectively. For anyone working in financial inclusion, public savings mobilisation, or India Post’s operations in cities like Coimbatore, these findings point to a clear and correctable problem. The post office has the trust, the products, and the physical infrastructure. What it is missing is communication that actually reaches people.

Awareness Of Post Office Savings Schemes Among The Public In Coimbatore: An Empirical Study On Knowledge Levels, Investment Behaviour, And Accessibility Barriers

Authors: Mohammed Anshaf.A, Dr. S.V. Harshini

Abstract: Walk into any post office in Coimbatore on a weekday morning and you will see the same thing: a queue of people waiting to pay electricity bills, pick up parcels, or send money home. Very few are there to ask about the Public Provident Fund or the Senior Citizen Savings Scheme. That is not because those products are bad — by almost any measure, they are excellent. It is because most people simply do not know they exist, or know only enough to be unsure where to start. This paper examines exactly that problem — the awareness gap around Post Office savings schemes — using survey data from 100 residents of Coimbatore district, collected between November 2024 and March 2025. We kept the analysis simple: structured questionnaire, simple percentage analysis, and four theoretical frameworks to make sense of the patterns. What we found was striking enough that we think it deserves a clear, direct write-up rather than one buried in statistical complexity. The headline number is this: 70% of respondents do not have full awareness of the Post Office savings schemes available to them. Word-of-mouth from family and friends is carrying most of the awareness load at 36%, while social media — which dominates how this city’s under-40 population discovers almost everything else — accounts for just 6% of scheme awareness. Safety and government backing motivate two-thirds of investment decisions. Service quality is where the experience lets people down most. And yet, 68% of those surveyed said they would recommend Post Office schemes to someone they know. There is genuine goodwill here. It is just not being channelled effectively. For anyone working in financial inclusion, public savings mobilisation, or India Post’s operations in cities like Coimbatore, these findings point to a clear and correctable problem. The post office has the trust, the products, and the physical infrastructure. What it is missing is communication that actually reaches people.

Patient-Centric Pharmaceutical Formulation Strategies For Paediatric And Geriatric Populations: Challenges, Regulatory Perspectives, And Emerging Technologies

Authors: Dr. Sabu M C, Sujith M, Dr. PriyaThambi T, Dr. Remya Krishnan G R

Abstract: Patient-centric pharmaceutical formulation has emerged as a critical paradigm in modern drug development, driven by the need to address the distinct physiological, functional, and behavioral characteristics of vulnerable populations such as paediatric and geriatric patients. Conventional, adult-centric dosage forms often fail to meet the safety, efficacy, and usability requirements of these age groups, leading to suboptimal therapeutic outcomes and poor adherence. Paediatric populations exhibit dynamic developmental changes in absorption, distribution, metabolism, and excretion, alongside heightened sensitivity to excipients and formulation attributes such as taste and dosing flexibility. In contrast, geriatric patients experience age-related organ function decline, polypharmacy, dysphagia, and cognitive impairment, all of which complicate medication administration and long-term adherence. This review critically examines patient-centric formulation strategies tailored to paediatric and geriatric populations, integrating physiological and biopharmaceutical considerations with practical formulation design challenges. Comparative evaluation of oral, parenteral, topical, and transdermal delivery systems is presented, highlighting how dosage form selection, excipient choice, and device design influence acceptability, safety, and therapeutic consistency across age groups. Regulatory perspectives from major agencies, including the US Food and Drug Administration and the European Medicines Agency, are analyzed to contextualize evolving expectations for age-appropriate formulations, excipient justification, and human factors engineering. Beyond established approaches, the review explores emerging technologies such as three-dimensional printing, nanocarrier-based systems, digital adherence tools, and smart drug delivery devices as enablers of personalized and adaptive therapy. By synthesizing current evidence, regulatory guidance, and technological advances, this article identifies key knowledge gaps and future research priorities essential for advancing truly patient-centric pharmaceutical design. Overall, the review underscores the necessity of integrating scientific rigor with real-world usability to optimize medication outcomes in paediatric and geriatric care.

AI-Based Pharmaceutical Supply Chain Optimization Using Prophet, LSTM, And Agentic Workflows

Authors: Keerthan Muni Raja T, Sanjay G S, Yokesh G A, Mrs.S.Deepa

Abstract: Public hospital medicine supply chains face persistent challenges including stockouts, medicine expiry, delayed procurement, and inefficient manual inventory control. While lean management principles provide structured waste reduction methodologies such as Just-In-Time (JIT) and Kanban systems, they lack predictive intelligence and real-time adaptability required in dynamic healthcare environments. This research proposes an AI-Guided Lean Decision Support System integrating machine learning-based demand forecasting, intelligent inventory monitoring, and lean rule enforcement to enhance operational efficiency. The proposed system architecture combines predictive analytics, real-time stock evaluation, automated reorder calculations, and route optimization to create a responsive and resilient medicines supply chain. The framework aims to reduce wastage, improve availability, and enable data-driven decision-making in public hospitals.

DOI: https://doi.org/10.5281/zenodo.19000626

GAN BASED FRAMEWORK OF DECISION SUPPORT FOR STOCK MARKET FORECASTING

Authors: Ms.Nirmala.S, Naasim.M, Harish.M, Aravamuthan.J.C, Aditya Raajan.M

Abstract: Stock market forecasting is a complex and challenging task due to the highly dynamic, nonlinear, and unpredictable nature of financial markets. Investors and traders continuously seek reliable tools that can help them make informed decisions while minimizing financial risk. Traditional statistical and machine learning approaches often struggle to capture sudden market fluctuations and hidden patterns present in stock price movements. To address these limitations, this paper presents a GAN-based framework for decision support in stock market forecasting. The proposed framework employs Generative Adversarial Networks (GANs) to learn complex market behavior from historical stock data and technical indicators. The generator network predicts future stock price movements, while the discriminator network evaluates the realism of these predictions by comparing them with actual market data. Through this adversarial learning process, the system improves forecasting accuracy and robustness. In addition to prediction, the framework provides decision support in the form of trend interpretation and buy–sell–hold signals, assisting users in practical investment decision-making. The proposed system aims to offer a reliable, intelligent, and data-driven solution for stock market forecasting.

DOI: https://doi.org/10.5281/zenodo.19001781

Smart Incident Assistant

Authors: V. Latha Sivasankari, Prijitha S, Bhavani B, Prem N

Abstract: Cloud computing infrastructures generate extensive system logs that record operational activities, performance metrics, and error events. While these logs are essential for monitoring system health, they are often technical, lengthy, and difficult to interpret quickly. The Smart Incident Assistant is a Java-based desktop application developed to transform raw cloud incident logs into structured, meaningful insights. This system acts as an intelligent incident summarization and recommendation platform, building upon a prior cloud log analyzer system. The application reads incident data from JSON files, processes the content using rule-based logic, classifies incidents based on severity levels, and generates human-readable summaries. Furthermore, it recommends suitable troubleshooting steps for each identified issue. The application is implemented using Java 11 and JavaFX, providing a clean graphical dashboard that allows users to monitor, manage, and export incident reports efficiently. By reducing manual log interpretation effort and improving response time, the Smart Incident Assistant enhances operational decision-making in cloud environments. The system demonstrates effective utilization of object-oriented programming principles, modular architecture design, file handling mechanisms, and intelligent rule-based processing techniques.

DOI: https://doi.org/10.5281/zenodo.19027776

AI Based Self Driving Cars

Authors: Smt. Preethi H U, Ms. Pallavi D G

Abstract: Artificial Intelligence (AI) has evolved into a defining technology of the modern era, profoundly transforming the global automotive industry. One of its most remarkable applications is the development of autonomous or self-driving vehicles that can sense their surroundings and navigate without human intervention. These vehicles integrate advanced AI models, machine learning (ML), and deep learning (DL) algorithms to process environmental data, recognize patterns, and make real-time driving decisions. The objective of this paper is to present a comprehensive survey of AI-driven self-driving car technologies, their system architecture, sensor networks, perception models, and control mechanisms. The study explores the evolution of intelligent transportation systems, emphasizing computer vision, sensor fusion, neural networks, and decision-making frameworks. Additionally, it examines the ethical, legal, and social challenges associated with autonomous vehicles, as well as future research opportunities. This paper concludes that AI-powered self-driving cars hold the potential to revolutionize mobility, enhance road safety, and contribute to sustainable urban development.

Intelligent Speech Fluency And Sign Translation System

Authors: Mr. R Madanachitran, Adaikkammai R, Mayuri P, Nandhini S S

Abstract: The Intelligent Speech Fluency and Sign Translation System is designed to enhance communication accessibility and speech development through advanced artificial intelligence techniques. The system integrates speech recognition, natural language processing, and computer vision to analyze spoken language and translate it into corresponding sign language gestures in real time. It also evaluates speech fluency by detecting pronunciation errors, pauses, repetitions, articulation issues, and speech rate irregularities. A deep learning–based acoustic model processes audio signals, while a convolutional neural network interprets sign gestures for bidirectional communication. The proposed framework supports inclusive interaction between hearing-impaired individuals and fluent speakers, promoting social integration and educational support. The system architecture ensures low latency and high accuracy through optimized preprocessing, feature extraction, and classification modules. Experimental evaluation demonstrates improved fluency assessment accuracy and reliable sign translation performance under diverse environmental conditions. This solution can be deployed in educational institutions, healthcare centers, and public service environments to foster inclusive and intelligent communication.

DOI: https://doi.org/10.5281/zenodo.19017430

Strategic Sales And Marketing Practices In The Pharmaceutical Industry: An Extended Research Study

Authors: Shadan Khan, Sana Moid

Abstract: The pharmaceutical industry plays a vital role in global healthcare systems. This research paper analyzes strategic sales and marketing practices within the pharmaceutical sector, with special emphasis on the role of digital marketing, relationship marketing, and professional promotion through healthcare practitioners. The study explores how pharmaceutical firms communicate product value, build relationships with physicians, and adapt to regulatory environments. Data and conceptual insights are drawn from internship observations in the pharmaceutical marketing domain and secondary literature. The findings indicate that physician engagement, ethical marketing practices, digital promotion, and efficient distribution channels significantly influence pharmaceutical product adoption. The paper concludes that pharmaceutical companies must balance ethical responsibility with competitive marketing strategies while embracing digital transformation to remain effective in modern healthcare markets.

DOI: https://doi.org/10.5281/zenodo.19017706

Multi-Modal Deep Learning Models For Image And Text Integration

Authors: Mr.K.Ajay Rathnavel, Ms.M.Vivitha, Mr.K.Rithik, Ms. B.Vinitha

Abstract: The integration of different modalities in deep learning models facilitates the incorporation of various data forms such as images and text, which enhances the multi-modal task performance. These models tackle issues such as representation of features, alignment of modalities, and strategies for fusion. The state of the art utilizes contrastive architectures such as CLIP, ALIGN, vision-language transformers like ViLT and Flamingo, and other hybrid components to boost cross-Moden reasoning. Tasks include image captioning, visual question answering, and multi-modal retrieval. Subsequent objectives combine architectures with efficiency in training and alignment techniques. Multi-modal learning is instrumental in pushing the boundaries of AI and its applications in comprehending the multifaceted nature of the real world.

DOI: https://doi.org/10.5281/zenodo.19019046

Ensemble Learning Approaches For Brain Stroke Detection

Authors: Dr Y. Subba Reddy, K. Uday Kiran, B. Gayathri, B.P. Anuhya Royal, K. Prasad

Abstract: A brain stroke is a medical emergency that occurs when the blood supply to a part of the brain is disturbed or reduced, which causes the brain cells in that area to die. The process involves training a machine learning model on a large labelled dataset to recognize patterns and anomalies associated with strokes. The proposed model is an ensemble machine learning algorithm, which integrates predictions obtained from several individual classifiers like Random Forest, Decision tree, KNN (K-nearest neighbor), voting Classifier, Logistic regression, XG Boost (Extreme Gradient Boosting) to make a final prediction. Each classifier provides a probability estimate for each class, final prediction is based on weighted average of these probabilities. The weights assigned to each classifier can be based on their performance on a validation set or can be set uniformly. The proposed voting classifier improved accuracy and robustness of final prediction compared to a single classifier. The limitation of study classification of stroke type can lead to appropriate use of resources and help to reduce healthcare costs. The proposed model obtained an accuracy of 100%. In order to carry out the investigation, the stroke prediction dataset is collected from UCI machine learning repository. The proposed system analyzes medical datasets containing patient attributes such as age, blood pressure, glucose level, heart disease history, and imaging data. Feature preprocessing and data balancing techniques are applied to improve model performance. The ensemble models are evaluated using performance metrics including accuracy, precision, recall, and F1-score.

Fake Social Media Profile Detection Using Machine Learning

Authors: Pooja Dharshini K

Abstract: Fake profiles are often created to spread misinformation, disinformation, and propaganda. Detecting and removing these profiles is crucial to curb the dissemination of false information and maintain the integrity of information shared on social media. In this study, we present an enhanced algorithm for the detection of fake social media profiles, utilizing machine learning techniques such as Gradient Boosting, Random Forest, and Support Vector Machine. The algorithm incorporates a range of profile features, including the presence of a profile picture, characteristics of the username and full name (length, numbers, equality), description length, external URL presence, account privacy, and key metrics like the number of posts, followers, and follows. The primary objective is to address the escalating issue of fraudulent activities and misinformation on social media platforms. The proposed algorithm leverages ensemble learning to improve the accuracy and reliability of identifying deceptive profiles. Additionally, we introduce a Flask-based web application to deploy the Random Forest algorithm, enabling real-time detection and providing a user-friendly interface. To evaluate the algorithm’s performance, precision, recall, and F1 score are employed as key metrics. Precision measures the accuracy of positive predictions, recall gauges the algorithm’s ability to capture all positive instances, and the F1 score balances precision and recall. Through comprehensive testing and validation, our algorithm aims to contribute to the advancement of online security, fostering user trust and mitigating the impact of fake profiles in the dynamic landscape of social media.

Real-Time Visual Assistance For The Visually Impaired Using Esp32-Cam

Authors: Mr. D. Nagaraju, K. Balaswamireddy, V. Jaya Vishnu, B.Balaji, U.Ganesh

Abstract: IoT based smart glasses system is designed for real-time environmental assistance and navigation for visually impaired people. The system is implemented using an esp32 microcontroller, integrating a camera module and a microphone for accurate scene capture and voice command processing. The camera continuously monitors the environment and converts visual signals into digital data from precise processing. Object, text and obstacles by analyzing the captured images through a cloud based AI platform. The esp32-cam uploads image and voice prompt data to the cloud platform through WiFi. Alerts recording detected objects, read text, and sense descriptions are sent to the user through an audio speaker using text-to-speech technology. The smart assistance system enhances user safety, improves navigation and provides independent living capabilities, making it cost effective and reliable solution for modern visually impaired users.

DOI: https://doi.org/10.5281/zenodo.19044359

Emotion Detection In Animal Vocalization Using Machine Learning Techniques

Authors: Ms. Vaishnavi Tandel, Ms. Malvika Thakur, Ms. Payal Nirmal, Mrs. Akanksha Patil

Abstract: An innovative AI powered web application which offers pet owners valuable insights into their animals’ emotional well-being and decodes pet emotions through audio analysis. The system processes audio recordings to classify emotions into categories such as happy, sad, anxious and fearful using advanced machine learning models, and also tracks behavioral patterns over time. An intelligent health agent identifies possible health problems or behavioral issues, offering personalized advice. This is based on a pet’s breed, age, and gender. It works for both dogs and cats, and includes secure user management and profiles specific to each breed.This enhances pet- owner bonding and enables early intervention for medical or behavioral problems, thus reducing the gap between human understanding and animal communication.

Deep Learning for Alzheimer’s Detection: A Smart Approach to Early Diagnosis

Authors: Haripriya T, Dr. D. Parameswari

Abstract: Alzheimer’s Disease (AD) is a progressive neurodegenerative disorder that primarily affects memory, cognitive function, and behavior, leading to severe impairment in daily activities. It is one of the most common causes of dementia among the elderly population worldwide and poses a significant burden on healthcare systems and society. Early diagnosis of Alzheimer’s disease is essential for timely intervention, effective treatment planning, and slowing disease progression. However, conventional diagnostic techniques rely heavily on neuroimaging interpretation and neuropsychological assessments, which are often time-consuming, expensive, and dependent on clinical expertise. Recent advances in deep learning (DL) have demonstrated remarkable potential in automating the diagnosis of Alzheimer’s disease using medical imaging data. This paper presents a de- tailed analysis of deep learning-based techniques for Alzheimer’s disease detection, with a particular focus on convolutional neural network (CNN) architectures applied to magnetic resonance imaging (MRI) and non-MRI modalities. In addition to the analytical review, this study implements a CNN-based Alzheimer’s disease detection system capable of classifying brain MRI images into four clinical stages: Non-Demented, Very Mild Demented, Mild Demented, and Moderate Demented. The proposed end-to- end framework enables automatic feature extraction and multi- class classification without the need for handcrafted features. Experimental observations demonstrate the feasibility and scalability of CNN-based approaches for early Alzheimer’s disease detection and their potential application in clinical decision support systems.

DOI:

Effect of Land Use Changes and Soil Depth on Selective Soil Physico-Chemical Properties in Tarn Taran District of Punjab

Authors: Harmandeep Singh, Manpreet Singh, Saurabh Sharma, Mamta Devi, Samrath Singh

Abstract: Soil degradation and changes in physico-chemical properties significantly impact crop production, posing challenges for feeding the growing population. This study evaluated these properties under various land-use systems in Tarn Taran district, Punjab. Soil properties such as bulk density (1.34–1.48 g cm⁻³), porosity (33.85–50.21 %), water holding capacity (30.47–46.61 %), pH (7.44–8.58), electrical conductivity (0.19–0.42 dS m⁻¹), soil organic carbon (0.42–0.69 %), and available nutrients (nitrogen: 235.45–271.29 Kg ha⁻¹, phosphorus: 18.72–26.03 Kg ha⁻¹, potassium: 183.39–363.47 Kg ha⁻¹) were analyzed at two depths (0-20 and 20-40 cm) across five system systems i.e., pear orchard, fodder-based, legume-based, rice-wheat, and sugarcane system. The results indicated that the pear orchard system exhibits the best soil health, with higher organic carbon, nitrogen, and potassium levels, likely due to organic management practices. In contrast, the sugarcane system shows the lowest nutrient availability and soil health, reflecting intensive nutrient removal and poor replenishment. Nutrient content generally decreased with soil depth, with surface soils being more enriched. This study underscores the need for sustainable soil management practices to improve nutrient cycling, prevent soil degradation, and enhance long-term agricultural productivity and ecosystem resilience.

DOI: https://doi.org/10.5281/zenodo.19046457

Real-Time Interactive 3D Particle System Using Three.js And Hand Gesture Control

Authors: Anjali Panchal, Sakshi Dhage, Priya Ingale, Manisha Bhadke

Abstract: This research paper presents the development of a real-time interactive 3D particle system that can be controlled using hand gestures captured through a camera. The system is implemented using Three.js to render interactive 3D graphics within a web browser. Gesture recognition techniques are used to detect hand movements and convert them into commands that control particle behavior. The system allows users to dynamically interact with particles by changing their size, color, and shape in real time. Various particle templates such as hearts, flowers, Saturn rings, and fireworks are implemented to enhance visual interaction. The integration of gesture recognition and 3D visualization improves human–computer interaction and creates an immersive experience. This project demonstrates the potential of combining artificial intelligence concepts with web-based graphics technologies for interactive applications in digital art, gaming, education, and visualization systems.

AI Tools in Investment Decision-Making: A Systematic Review of Applications, Challenges, and Opportunities in India

Authors: Fayez Ameer Kozhithodi, Fathima Salna K, Mubeena Valiyapeediyekkal, Aswathi ES

Abstract: Artificial Intelligence (AI) tools, such as robo-advisors, predictive analytics, and risk management systems, are transforming investment decision-making in India’s rapidly expanding financial markets. With over 120 million retail investors participating in the National Stock Exchange (NSE) and Bombay Stock Exchange (BSE), AI-driven platforms play a critical role in enhancing decision accuracy, mitigating behavioral biases, and improving market efficiency. This systematic literature review follows PRISMA guidelines, synthesizes 16 selected sources, primarily peer-reviewed articles and authoritative reports published between 2020 and 2026. The review evaluates AI applications, empirical performance outcomes—including predictive accuracy improvements of 20–30% in several emerging-market studies, high investor acceptance (mean scores around 4.2 approximately on five-point Likert scales from TAM-based surveys), and notable enhancements in risk management and fraud detection through AI tools—and persistent challenges related to regulation and data privacy under the Digital Personal Data Protection (DPDP) Act, 2023. Thematic analysis following Braun and Clarke (2006) identifies three dominant clusters: LSTM-based Nifty forecasting models, robo-advisory platforms integrated with UPI systems, and real-time AI-driven risk management frameworks contributing to improved operational efficiency and market surveillance. The findings highlight India’s rapid FinTech adoption relative to global trends while emphasizing concerns related to algorithmic opacity. The study extends the Technology Acceptance Model (TAM) by incorporating data privacy as a moderating factor and advocates hybrid human–AI decision frameworks. Policy, educational, and research recommendations are proposed to support sustainable and explainable AI adoption in Indian capital markets.

DOI: https://doi.org/10.5281/zenodo.19049218

A Review On IOT Wireless Network Optimization Models And Techniques

Authors: Roshni Thakre, Prof. Rahul Patidar, Jayshree Boaddh

Abstract: Wireless Sensor Networks (WSNs) and Internet of Things (IoT) systems play a crucial role in modern smart applications such as smart cities, healthcare, industrial automation, and precision agriculture. Despite rapid advances in communication, cloud, and edge computing technologies, energy efficiency remains a fundamental challenge due to the limited power resources of sensor nodes. This survey presents a comprehensive review of energy losses in WSN-based IoT networks and systematically analyzes state-of-the-art optimization techniques aimed at prolonging network operation. The paper discusses architectural aspects of WSNs, sources of energy waste, and major optimization strategies, including radio optimization, data reduction, sleep/wakeup scheduling, and energy-aware routing. Furthermore, recent approaches based on metaheuristic optimization, swarm intelligence, and machine learning are reviewed and compared. By organizing existing methods into a unified taxonomy, this survey highlights current research trends, practical design trade-offs, and open challenges, thereby providing useful guidelines for developing scalable, energy-efficient, and reliable WSN-IoT systems.

Feminist Voices In Modern Tamil Literary Works

Authors: Ms.R.Vilva Viruksha, Dr.R.Shiva Shankari, Ms.S.Chitra

Abstract: This paper undertakes a critical analysis of the role of feminism in contemporary Tamil literature, tracing the development of women’s expression, resistance, and agency in the late twentieth and twenty-first centuries. Through the critical analysis of foundational Dalit feminist literature, confessional poetry, and the use of narrative forms in contemporary literature, the paper identifies three significant stages in the development of women’s expression and agency in Tamil literature. This study proposes the use of Intersectional Narrative Analysis (INA) as a methodological framework to analyze the role of women writers in Tamil literature, which combines the analysis of the narrative with the socio-political context. Through the analysis of the works of Bama, Salma, Kavitha Sornavalli, and S. Ramakrishnan’s novel *Idakkai*, the study identifies the evolution of women’s literature in Tamil, which has moved from the narrative of victimhood to the expression of autonomous subjectivity, bodily autonomy, and the reconstruction of the social order. Through the evaluation of the works of women writers on five dimensions—narrative, themes, resistance, intersectionality, and impact—this study identifies the unique contribution of each work and the common thread of the expression of women’s agency and resistance to patriarchal hegemony. This indicates that modern feminist writing in Tamil is using experimental narrative techniques to convey the complexity of women’s experiences while engaging with broader social justice movements. The present research is significant in terms of exploring regional literary traditions and their engagement with global feminist discourses and cultural specificity.

DOI: https://doi.org/10.5281/zenodo.19049707

An Analytical Review Of Thought Simulation In Artificial Intelligence Tools

Authors: Dr. D. Muthulakshmi, S. Abinaya, B. Kaviya

Abstract: Simulation of human thought processes using artificial intelligence is seen as a revolutionary approach toward the design and development of cognitive computing paradigms, moving beyond pattern recognition and matching architecture. This paper discusses a thorough analytical review of thought simulation methodologies used in contemporary artificial intelligence systems, highlighting the transformation from large language models to cognitive artificial intelligence architecture. In this regard, through a thorough literature survey of 30 peer-reviewed articles published between 2021 and 2026, we identified three major thought simulation methodologies: personalized cognitive simulation using large-scale behavioral datasets, emulative cognitive simulation using world models for situated reasoning, and developmental cognitive architecture using functional equivalence to human cognitive growth. We propose a unified methodological framework for cognitive simulation architecture (CSA). The analysis process will be accompanied by five representative figures on the comparison of simulation fidelity, components of the architecture, performance measures, and comparative evaluation with reference models. A comparative analysis table will be included, which will compare existing approaches on seven criteria: cognitive grounding, scalability, transparency of reasoning, consistency of interventions, and computational efficiency. Our analysis suggests that the development of thought simulation capabilities has moved beyond the simple mimicry of surface-level behavioral processes and has reached the development of architectures with internal coherence, traceability, and the emergence of collective intelligence. However, many challenges remain to be overcome in the development of functional equivalence with human cognition and the ethics of the process.

DOI: https://doi.org/10.5281/zenodo.19049937

Study on Artificial Intelligence in Human Resource Management

Authors: Bindu H T, Yashaswini A, Shilpa Raj, Archana M

Abstract: Human resource management (HRM) has experienced a significant change due to the groundbreaking advancements introduced by artificial intelligence (AI). This study investigates the transformation of HRM in an AI- centric landscape, highlighting the shift from traditional approaches to data- informed, technologically sophisticated decision- making processes. Integrating AI into HR functions has enhanced efficiency and customization by optimizing hiring, training, performance evaluations, and employee involvement. However, these advancements also raise organizational, social, and ethical issues.

DOI: https://doi.org/10.5281/zenodo.19050116

Learn2Play: An AI-Integrated Educational Ecosystem For Game Development Students

Authors: Sushmita Kirar, Sanskruti Wagadare, A. P. Bhosale

Abstract: – The rapid growth of the global game development industry has created an increasing demand for skilled professionals with both technical and creative expertise. Traditional teaching approaches often fail to provide the interactive and practical experiences required for mastering game development skills. This paper introduces Learn2Play, an Android-based educational ecosystem designed specifically for students learning game development. The platform integrates Firebase Authentication, Cloud Firestore, and Realtime Database to deliver a secure and scalable learning environment. Learn2Play includes multiple modules such as AI-powered learning assistance, gamified learning activities, video lectures, assignment management, and real-time progress analytics. The system is developed using Java and XML while following Material Design 3 design principles to ensure usability and consistency. Security features such as Firebase App Check and session management through Shared Preferences are incorporated to ensure reliable access and user data protection. Results from pilot deployment involving 120 students indicate increased engagement, improved learning outcomes, and higher assignment completion rates compared with traditional teaching methods. The findings demonstrate the effectiveness of mobile-first learning systems combined with artificial intelligence for technical education.

SmartC Professor – AI-Powered Learning Companion For C Programming

Authors: Vivek Nagargoje, Neha Sawakar, Om Kadam, Aryan Marathe

Abstract: Artificial Intelligence (AI) has transformed modern education by enabling adaptive and personalized learning experiences. Programming, one of the most essential skills in computer science, continues to pose challenges to beginners due to the technical nature of compiler errors and complex syntax understanding. Traditional compilers detect errors but fail to explain them in learner-friendly language. To address this issue, SmartC Professor is proposed — an AI-driven tutoring system designed to enhance learning in C programming. The system integrates compiler analysis with Natural Language Processing (NLP) to generate intelligent, human-readable feedback and explanations. It also features an AI chatbot that answers conceptual C programming questions. Implemented using Python’s Flask framework, GCC compiler, and NLP algorithms like TF–IDF and Cosine Similarity, the system provides real-time guidance to students. Results demonstrate that SmartC Professor improves students’ understanding, reduces debugging time, and fosters independent learning.

Factors Influencing Ride Height In Vehicle Dynamics

Authors: Raghunath Parate, Priyanka Ranalkar, Prajwal Bhadange

Abstract: This paper examines the factors influencing vehicle ride height, focusing on elements that affect dynamics, performance, and stability. These factors are categorized into four groups: Man (Human/Assembly Factors), Material (Component Factors), Method (Operational Factors), and Machine (Tools and Equipment). Human errors, such as improper assembly, can directly affect ride height. Material issues, like inappropriate spring or rubber seat stiffness, also play a key role. Operational factors, including overloading or incorrect height measurement, contribute further. Lastly, improper tools can lead to inaccurate readings. Understanding these factors is crucial for maintaining optimal ride height, which enhances vehicle safety, comfort, and performance.

Virtual Prediction Of Clutch Disc Misalignment In Powertrain Assembly Across Its Design Tolerance Range

Authors: Pranay Subhedar, Omkar Patil, Nitin Tawhare

Abstract: Clutch-disc alignment plays a critical role in the NVH performance, durability, and shift quality of automotive drivetrains. Even minor angular or lateral misalignments between the clutch disc, flywheel, and transmission input shaft can generate harmful dynamic loads, accelerated spline wear, and undesirable vibration signatures. Traditional alignment evaluation relies heavily on physical prototyping and end-of-line inspection, which are often reactive, cost-intensive, and limited in their ability to isolate root causes during early design stages. This paper presents a virtual methodology for defining, predicting, and validating clutch-disc misalignment criteria using Dimensional variation analysis (DVA) as a design-for-quality tool. The approach establishes quantifiable misalignment thresholds based on dynamic system behavior, correlates them with component tolerances, and evaluates their influence on clutch engagement quality and torsional response. A comprehensive DVA-based workflow is proposed for early-phase design optimization, enabling engineers to identify misalignment-induced excitation modes, assess factor sensitivities, and ensure robust manufacturability across platform variants. The results demonstrate how virtual misalignment analysis significantly mitigates prototype iterations, improves alignment robustness, and enhances system quality for diverse automotive applications.

Ai-Enhanced Real Time Speech To Speech Translation

Authors: Barath Raaj S A, D.Parameswari, Varnikhasri S, Udaya Kumar S

Abstract: The demand for effective communication through practical means continues to increase due to globalization. In such areas as education, health care, tourism and multinational cooperation, everyone faces some type of language barrier; therefore, all of these situations rely on an effective communication system to facilitate dialogue between individuals who speak multiple languages in their native language. This paper discusses a system that allows people who speak different languages to interact with each other in real time using a real-time speech-to-speech system (S2ST). It consists of three major parts: automatic speech recognition (ASR), neural machine translation (NMT) and text to speech (TTS) synthesis. ASR is the part of the system that takes the spoken input and converts it into an electronic form of the input as soon as the speaker has finished speaking. ASR uses streaming ASR technology to produce this conversion in near real time while removing background noise using noise reduction and detecting when there is actual voice activity so that the end user receives accurate and robust information under all conditions. The NMT portion of the S2ST system employs a transformer based neural translation model to generate translated electronic forms of the output of the ASR component. The NMT component uses semantic meaning and contextual correctness rather than simply using the word-for-word translations for the output of the ASR. The TTS portion of the S2ST system produces quality, natural sound outputs in order to enable conversational interaction. The experimental study of this system confirms that the system produces high levels of accuracy in translation and barriers and promoting inclusive global communication.

DOI: https://doi.org/10.5281/zenodo.19087958

Open IMScore Security Enhancedcwith Private Cloud

Authors: Abdallah Handoura

Abstract: The Next Generation Network (NGN) is an IP-based, packet-oriented telecommunications architecture designed to support a wide range of services. The IP Multimedia Subsystem (IMS), in comparison, provides the control layer and service delivery framework that enables multimedia applications to function over NGN. While NGN constitutes the core network infrastructure, IMS serves as the specialized platform responsible for delivering and managing multimedia services such as voice and video communications to end users. Security for users, services, and providers is a fundamental requirement. The growing sophistication of security threats, combined with service diversity and the widespread adoption of cloud-based and distributed systems, has made comprehensive security a primary concern. In this work, a holistic security mechanism is introduced to improve protection across the IMS environment by integrating IMS architecture with cloud computing platforms. This integration leverages the cloud’s advanced security capabilities, multiple deldiveeferynseframlayeewros,rkanthdatcoanlltorwols mopeecrhaatnoirssmtso tooffer strengthen system reliability, confidentiality, and overalaldrovabnucsetndeNssG. N services. As shown in Figure 1, the This paper presents an integration between two fields: telecommunication services and network concepts along with open technologies, aiming to provide an open service creation framework and to leverage the security mechanisms offered by cloud-based authentication.

DOI: https://doi.org/10.5281/zenodo.19088856

Enhancing Transparency In E-Commerce Recommendations Through Feature-Enhanced Natural Language Explanations (FENLE)

Authors: Sahabdeen Aysha Asra

Abstract: E-commerce recommendation systems often function as opaque “black boxes,” limiting user trust and engagement despite their predictive accuracy. This concept paper proposes Feature-Enhanced Natural Language Explanations (FENLE) to address this challenge by combining interpretable product features with dynamic, user-friendly natural-language justifications. A structured methodology is outlined for designing and testing FENLE in experimental settings, with simulated results suggesting significant improvements in user trust, satisfaction, perceived usefulness, and engagement compared to baseline and simple feature-based explanations. While acknowledging trade-offs between accuracy and explainability, the paper emphasizes the value of user-centered design and adaptive explanation strategies. By integrating Explainable Artificial Intelligence (XAI) principles into recommender systems, e- commerce platforms can foster transparency, loyalty, and responsible AI adoption, aligning technical innovation with user expectations and ethical standards.

DOI: https://doi.org/10.5281/zenodo.19089652

Automating Libraries: The Role Of Library Management Systems In Modern Library Operations

Authors: M.V.Subba Reddy Professor, A.Sana Tabassum, D.Karthik Reddy, I.P.Sruthi, K.Sandeep Reddy

Abstract: This study examines the significant role of LMS within the context of modern library functionality, emphasizing its influence on operational efficiency through an extensive review of the literature derived from esteemed academic databases, such as Scopus, Web of Science, Taylor & Francis, and Google Scholar. The findings of this research elucidate the various applications of LMS, which encompass cataloging, circulation management, acquisitions, integration of digital resources, and data-driven decision-making. Moreover, the study identifies critical challenges that hinder the effective implementation of LMS, including technical constraints, organizational resistance, and human-related factors, such as insufficient staff training. To mitigate these issues, the research advocates for strategic approaches that incorporate comprehensive needs assessments, continuous professional development initiatives, the selection of flexible technology solutions, adherence to user-centered design principles, robust data security measures, and ongoing evaluation processes. In conclusion, this research underscores the dynamic nature of LMS, influenced by technological advancements such as artificial intelligence (AI), radio-frequency identification (RFID), and big data analytics, which play pivotal roles in shaping the future of knowledge management.

DOI: https://doi.org/10.5281/zenodo.19091845

Effect Of Rear Wheel Steering On Turning Circle Diameter In Virtual Analysis

Authors: Raghunath Parate, Abhishek Atal, Priyanka Ranalkar

Abstract: Turning Circle Diameter (TCD) is a critical parameter in vehicle manoeuvrability, particularly in urban environments where tight turns and compact parking spaces are prevalent. Traditional front-wheel steering systems limit the potential for reducing TCD due to geometric and packaging constraints. This paper investigates the implementation of rear-wheel steering (RWS) as a means to reduce TCD. Based on simulation and theoretical analysis, two RWS configurations — 5° and 7° rear steer angles are evaluated. Results show an approximate 8% and 12% reduction in TCD, respectively. However, mechanical constraints such as wheel envelope clearance, interference with suspension components, and vehicle loading conditions must be addressed to ensure feasibility. The study concludes that RWS can be a highly effective solution for improving low-speed manoeuvrability when integrated carefully with physical constraints.

A Comparative Study of Some Methods of Estimating Parameters of Linear Regression in Presence of Multicollinearity and Outlier

Authors: Warha, Abdulhamid Audu, Akeyede, Imam, Grema Modu Bako

Abstract: Classical least squares method for estimating regression models consisting of minimizing the sum of the squared residuals. Some of the assumptions of Ordinary least squares method (OLS) is that there is no correlations (multicollinearity) and extreme values (outliers) between the independent variables. Violation of these assumptions arises most often in regression analysis and can lead to inefficiency of the least square method. This paper therefore determined the efficient estimator between Least Absolute Deviation (LAD) and Weighted Least Square (WLS) in multiple linear regression models at different levels of multicollinearity and outlier in the explanatory variables. Simulation technique were conducted using R Statistical software, to investigate the performance of the two estimators under violation of assumptions of lack of multicollinearity and outliers. Their performances were compared at different sample sizes. Finite properties of estimators’ criteria namely, mean absolute error, absolute bias and mean squared error were used for comparing the methods. The best estimator was selected based on minimum value of these criteria at a specified level of multicollinearity, outlier and sample size. The results showed that, LAD was the best at different levels of multicollinearity and outlier and was recommended as alternative to OLS under this condition. The performances of the two estimators decreased when the levels of multicollinearity and outliers was increased.

DOI: https://doi.org/10.5281/zenodo.19098391

Self-Compacting Concrete With Mineral Admixtures: An Experimental Study

Authors: Mora Raja Kumar, Dr.K.Naga Sreenivasa Rao

Abstract: Self-Compacting Concrete (SCC) has emerged as an advanced construction material due to its ability to flow under its own weight, completely fill formwork, and achieve full compaction without the need for external vibration. This experimental study investigates the influence of selected mineral admixtures on the fresh and hardened properties of SCC, aiming to enhance its performance and sustainability. Mineral admixtures such as fly ash, ground granulated blast furnace slag (GGBS), and silica fume were used as partial replacements of cement at varying proportions. The fresh properties of SCC were evaluated using standard tests including slump flow, T₅₀ time, V-funnel, and L-box tests to assess filling ability, passing ability, and segregation resistance. Hardened properties were examined through compressive strength, split tensile strength, and flexural strength tests at different curing ages. The results indicate that the incorporation of mineral admixtures significantly improves the workability, flow characteristics, and long-term strength of SCC while reducing cement content and heat of hydration. Among the mixtures studied, SCC incorporating an optimum combination of mineral admixtures exhibited superior performance compared to conventional SCC. The study confirms that the effective use of mineral admixtures in SCC not only enhances mechanical and rheological properties but also contributes to sustainable and eco-friendly concrete production.

Performance Of Recycled Aggregate Concrete For Structural Applications

Authors: Mutha Ajay, U. Srinivasarao

Abstract: Now-a-days cost of construction materials are affecting the economy of all structures. It is a dominating factor affecting environmental housing system around the world. Conventional aggregates namely gravel, and fine aggregate is sand in concrete will be used to control. While natural aggregate is Recycled Aggregate as coarse aggregate will be investigated to replace the aggregate in concrete and Robo sand (Stone dust) as fine aggregate will be replace the sand in concrete. In this investigation, M25 grade of concrete with combination of natural material Recycled Aggregate content as Coarse aggregate in the proportion of 0%,5%,10%,15%,20% & 25% will be replaced and Robo sand (stone dust) as fine aggregate with full 100% replacement of natural sand, sample specimens are prepared and will be tested for workability, compressive strength, split tensile strength and flexural strength for 7,14 and 28 days respectively and also showing the comparative results with Conventional M25 grade concrete. By this project investigation, concrete may be less dense, light weight concrete by Recycled Aggregates and good quality of concrete by Robo sand.

Bridge Load Rating And Analysis Using MIDAS Civil

Authors: Namathoti Snehalatha, G. Nagalakshimi

Abstract: The present effort focuses on improving the technique for analyzing and designing flexible pavement by taking into account the wide range of materials that will be utilized in the various pavement layers and the real field’s environmental conditions. In order to understand how different loading conditions and material qualities affect performance parameters, the 2D axis symmetric finite element approach is used. The horizontal and vertical limits of the mesh are established after assuming and validating the planned pavement sections from the relevant codes. Subgrade soils were categorized according to their strength parameters by field study and laboratory examination. This research analyzes the impact of various waste products, both singly and in combination, on the physical characteristics, rutting, and fatigue behavior of pavement in a controlled laboratory setting. Structural models make use of a wide variety of input data, such as traffic characteristics, moduli of paving materials, etc. Distress models employ structural model outputs such as strains, stresses, and deflection, and their results are compared to fatigue and rutting requirements from the design handbook of flexible pavement (IRC37-2012). By comparing the estimated strains at the key sites with the allowed limits, the thicknesses of hypothetical circumstances created under different combinations of different materials and varied thicknesses may be determined. The majority of India’s flexible pavement is showing signs of early deterioration, much before the end of its expected lifespan. The results of an analysis of the relationship between overloading and inflated pressure show that the former is likely a contributing factor to the premature degradation of pavement. Overloading has a greater impact on pavements built on poor subgrade soils; hence, this problem must be addressed at the pavement design phase. Back-calculated elastic moduli bituminous mix designs, as suggested by temperature analysis, are likely to have the best possible depth, making them more cost-effective. The proposed design charts are more realistic and optimal since they take into account a number of aspects of diverse nature; these charts may be used to replace an existing pavement section in a way that is both environmentally friendly and fiscally responsible.

Use Of Waste Glass Powder As Partial Replacement Of Cement In Concrete

Authors: Narne Bhavani, Dr.K.Naga Sreenivasa Rao

Abstract: Concrete has occupied an important place in construction industry in the past few decades and it is used widely in all types of construction ranging from small buildings to large infrastructural dams or reservoirs. Cement is major ingredient of concrete. The cost of cement is increasing day by day due to its limited availability and large demand. At the same time the global warming is increasing day by day. In the present study an attempt been made on concrete and also an experimental investigation on the concrete using by replacing cement with FLYASH and GROUND GRANULATED BLASTFURNACE SLAG(GGFBS) to decrease the usage of cement as well as emission of concrete. Experimental studies were performed on plain cement concrete and replacement of cement with Fly ash is done. The main properties of flyash are fineness, specific gravity, chemical composition, carbon content etc. the fineness of flyash is important because it effects the workability of the concrete. Specific gravity does not directly effect concrete quality, it has value in identifying changes in other flyash characteristics. The variability of the chemical composition is checked regularly as a quality control measure. GGBS is a byproduct of the manufacturing of iron in the blast furnace can partially replace cement in concrete. GGBS hardens very slowly so it is generally used along with OPC in concrete. A typical combination of GGBS(70,80,90%) and fly ash is used in 10, 20 and 30%.

Early Age Properties Of High -Strength Concrete With Chemical Admixtures

Authors: Sayyad Khasim, Dr.K.Naga Sreenivasa Rao

Abstract: High-strength concrete (HSC) is widely used in modern construction for high-rise buildings, long-span bridges, and other infrastructure requiring superior mechanical and durability performance. The early age properties of HSC, including workability, setting time, and early strength development, play a critical role in ensuring the quality, placement efficiency, and structural integrity of concrete structures. The incorporation of chemical admixtures, such as high-range water-reducing agents (superplasticizers) and accelerators, has become a common practice to enhance the fresh and early age properties of concrete, particularly in mixes with low water-to-cement ratios and high cement content. This study investigates the influence of chemical admixtures on the early age properties of HSC, focusing on their impact on workability, setting characteristics, and early compressive and tensile strength development. Laboratory investigations were carried out using ordinary Portland cement, well-graded fine and coarse aggregates, and high-range water-reducing admixtures conforming to IS: 9103–1999. Concrete mixes were prepared with varying dosages of superplasticizers to assess the effect on workability, measured using the slump test, and flowability, measured using the compacting factor test. Early age strength development was evaluated by casting cube and cylindrical specimens and testing them for compressive and split tensile strength at 1, 3, and 7 days of curing. Additionally, the setting time of cement paste with and without admixtures was determined using Vicat apparatus in accordance with IS: 4031–1988. The experimental results indicate that the inclusion of chemical admixtures significantly improves the workability of HSC, allowing a reduction in water content without compromising fluidity. Superplasticizers were observed to maintain a uniform, cohesive mix, reduce segregation, and facilitate proper compaction at early ages. Early age strength results demonstrated that appropriate admixture dosages accelerate hydration and improve initial compressive and tensile strength, which is critical for formwork removal, prestressing operations, and early load application. Setting time measurements showed that admixtures can either retard or accelerate initial and final setting, depending on their chemical composition, allowing better control over placement and finishing operations. The study highlights the importance of optimizing admixture type and dosage to balance workability, early age strength development, and setting time for high-strength concrete applications. Overall, this research confirms that chemical admixtures are essential for enhancing the early age performance of high-strength concrete, contributing to improved constructability, structural reliability, and long-term durability in modern construction projects.

Performance Of Concrete With Plastic Waste As Fine Aggregate

Authors: Shaik Munwar, Sk. Abdulkareem

Abstract: Pavements form a critical component of the transportation infrastructure, providing a safe, smooth, and durable surface for the movement of vehicles ranging from two-wheelers to heavy-duty trucks. The performance of pavements directly influences transportation efficiency, road safety, vehicle operating costs, and the overall comfort of commuters. A well-constructed and maintained pavement ensures smooth traffic flow, reduces wear and tear on vehicles, and minimizes travel time, thereby supporting economic growth and societal mobility. Conversely, poor pavement conditions lead to discomfort, increased maintenance costs, accidents, and higher energy consumption. The deterioration of pavements is a multifaceted problem resulting from both mechanical and environmental stresses. Mechanically, pavements are subjected to repeated traffic loading, which ranges from light two-wheelers to heavy commercial vehicles such as trucks and buses. The intensity, frequency, and axle configuration of these loads significantly influence pavement performance. Over time, heavy loading can cause surface deformations, cracking, rutting, and fatigue failure, particularly in the surface course and base course, which bear the majority of the applied stresses. The surface course may develop rutting, potholes, and micro-cracks, while the base course can suffer from structural weakening and loss of load-bearing capacity if not properly designed and constructed. Environmental factors also play a critical role in pavement deterioration. Rainfall, flooding, and water infiltration can weaken the subgrade, erode materials, and cause pavement settlement or potholing. Seasonal temperature variations lead to expansion and contraction cycles, contributing to thermal cracking, while extreme events such as earthquakes and heavy storms can induce abrupt structural damage. The combined effect of traffic loading and environmental exposure accelerates pavement distress, reducing its service life and increasing the need for maintenance interventions. The interaction between mechanical loading and environmental effects makes pavement design a complex process. Modern pavement engineering emphasizes the use of high-performance materials, proper layered design, and durability-focused construction practices to mitigate the adverse effects of heavy traffic and environmental exposure. Techniques such as improved mix designs, geosynthetics reinforcement, drainage management, and periodic maintenance are essential to prolong pavement life.

Mechanical Performance Of Ultra-High Performance Concrete (UHPC)

Authors: Vaddeswaram Pranath Kumar, V.E.S.Mahendra Kumar

Abstract: Ultra-High Performance Concrete (UHPC) represents a major advancement in modern concrete technology, offering superior mechanical, durability, and structural performance compared to conventional and high-performance concretes. UHPC is characterized by its extremely high compressive strength, enhanced tensile strength, and improved ductility, achieved through a combination of optimized particle packing, very low water-to-cement ratios, and the inclusion of supplementary cementitious materials such as silica fume, fly ash, and ultra-fine powders. The incorporation of steel or synthetic fibers further enhances tensile strength, flexural performance, and post-cracking behavior. The mechanical performance of UHPC is largely influenced by its dense and nearly impermeable microstructure, which minimizes porosity and refines the interfacial transition zone (ITZ) between cement paste and aggregates. Compressive strengths of UHPC typically exceed 150 MPa, while tensile and flexural strengths are significantly higher than conventional concrete, making it ideal for applications requiring extreme load-bearing capacity, structural resilience, and long-term durability.

Effect Of Nano-Silica On High-Strength Concrete Properties

Authors: Sunkara Janardhan, N. Sriaknth

Abstract: The incorporation of nano-silica (NS) into high-strength concrete has emerged as an effective approach to enhancing both mechanical and durability properties of cement-based materials. Owing to its extremely fine particle size and high specific surface area, nano-silica significantly influences the hydration process and microstructural development of concrete. This study investigates the effect of varying nano-silica contents on the fresh, mechanical, and durability properties of high-strength concrete. Nano-silica acts as a highly reactive pozzolanic material, accelerating cement hydration and promoting the formation of additional calcium silicate hydrate (C–S–H) gel, which leads to a denser and more refined microstructure. As a result, notable improvements in compressive strength, tensile strength, and flexural strength are observed, particularly at early curing ages. The presence of nano-silica also enhances the interfacial transition zone between aggregates and the cement matrix, reducing microcracks and porosity. This refinement contributes to improved resistance against water permeability, chloride ion penetration, and chemical attack, thereby significantly improving the durability performance of high-strength concrete. Additionally, nano-silica reduces bleeding and segregation, improving the homogeneity of the mix, although careful control of dosage and dispersion is required to prevent agglomeration and excessive water demand. Experimental observations indicate that an optimum nano-silica content exists beyond which strength and workability may be adversely affected.

VERITAS: Evidence-Based Regulatory Intelligence For Automated Document Compliance Validation

Authors: Sahil Bagul, Ganesh Ghadge, Karan Dhokale, Dheeraj Patil

Abstract: The rise in data-focused work and stricter privacy rules has increased the need for dependable compliance management. Companies must show they follow rules like GDPR, HIPAA, and ISO 27001. Each has complex terms and overlapping control needs. Current audits mostly involve manual document review. Compliance staff compares policies, steps, and reports to legal terms. This takes time, lacks consistency, and can lead to mistakes. This paper presents VERITAS: Validation and Evidence-based Regulatory Intelligence for Transparent Audit Systems, an automated tool. It validates compliance at the clause level using semantic search and reasoning. The system handles company documents, turns text into semantic vectors, and pulls up relevant sections for each legal clause. It then uses rule-based reasoning to decide compliance with clear support. We conducted a study across datasets covering GDPR, HIPAA, and ISO 27001 policies. The tool validated a 10-page document in 26.8 seconds, with 91.2% retrieval precision and 87.5% reasoning accuracy. This beats manual audits in speed and reliability. The tool also gives reports with linked evidence, ensuring audit trails and understanding. By merging retrieval-based reasoning, clear decision paths, and scalable automation, VERITAS sets a base for regulatory assurance and compliance in data-driven businesses.

Injury patterns in sports-related knee and shoulder injuries in Wrestlers from Haryana Practicing in Government Academies

Authors: Dr. Anjli

Abstract: Wrestling is an Olympic sport for both men and women. It is so ancient since 708 B. C. It is a combat sports that results in various types of injuries that occur in sports. Wrestling and football, which involve frequent player-to-player contact, have a high risk of shoulder injuries. Furthermore, wrestling involves constant contact with the playing surface, exacerbating the condition According to research from the Center for damage and Policy, football and wrestling are the two sports with the highest risk of major damage to athletes. Aim of this study is to know about injury patterns in sports-related knee and shoulder injuries in wrestlers from Haryana practicing in Government Academies. Data has been collected through off line structured questionnaire of wrestlers and coaches of Government Academies of Haryana State. Sample size for the collection of the data has been collected from the total 551 respondents (wrestlers) on the basis of exclusion and inclusion criteria. In our study population, 40 out of 551 players (7.25%) who experienced shoulder injuries and 77 out of 551 (13.97%) who experienced Knee injuries. In our study, 7.25% of participants reported shoulder injuries.

DOI: https://doi.org/10.5281/zenodo.19099736

Virtual Classroom System with Attendance Tracking Andai- Based Lecture Summaries

Authors: Dr. P. Vara Prsad, K.V.Sai Jayanthi, M.Anjali, G.Ashok Kumar, L.Rahul

Abstract: The rapid growth of online education has transformed the modern learning environment. With the increasing use of digital platforms and internet connectivity, managing virtual classrooms efficiently has become an important challenge for educational institutions. Traditional methods of tracking attendance and preparing lecture notes require manual effort and consume valuable teaching time. This paper proposes an Artificial Intelligence (AI) based virtual classroom system that automates attendance tracking and generates lecture summaries automatically. The system utilizes face recognition technology to identify students and record attendance during online sessions. In addition, speech-to-text technology converts lecture audio into text, and Natural Language Processing (NLP) techniques generate concise summaries of lecture content. The proposed system reduces administrative workload, improves attendance accuracy, and provides students with summarized lecture notes for quick revision. By integrating AI technologies into virtual learning environments, the system enhances the efficiency and effectiveness of online education.

DOI: https://doi.org/10.5281/zenodo.19106354

An Android-Based Mobile Application For Efficient College Event Management: Design, Implementation And Evaluation

Authors: Professor Maske P. P., Gayatri Shendge, Vinay Bhavsar, Sushant Kshirsagar, Saurabh bhosale

Abstract: The management of college events traditionally involves fragmented communication channels, cumbersome manual registration processes, and difficulties in tracking participation and real-time coordination among multiple stakeholders. This paper presents the design and implementation of a comprehensive Androidbased Event Management System aimed at automating and simplifying these processes through a centralized, multi-role platform. The system features a multi-user architecture with distinct interfaces and functionalities tailored to three primary user roles: Administrators, Event Coordinators, and Participants. Key functionalities include secure Firebase Authentication, dynamic event listings with real-time synchronization, comprehensive event descriptions, in-app registration modules, and granular role-based access control (RBAC) for differentiated event management capabilities. The application is developed using Java for the Android platform and utilizes Google Firebase services, including Realtime Database for data persistence and authentication mechanisms. This paper elaborates on the system architecture, the implementation of core modules, the technologies employed, and evaluation metrics demonstrating improved efficiency in event management. The resulting application offers a user-friendly, scalable, and efficient solution for managing academic and extracurricular events in educational institutions[1][2].

Experimental Study On Self-Healing Concrete Using Bacterial Additives

Authors: Puppala Harish Kumar, N. Sriaknth

Abstract: The durability and service life of conventional concrete structures are significantly compromised by the formation of microcracks, which facilitate the ingress of water and aggressive agents. This experimental study investigates the feasibility and performance of self-healing concrete incorporating bacterial additives as a sustainable crack remediation technique. A ureolytic bacterial strain capable of inducing calcium carbonate precipitation was introduced into concrete mixes using a suitable carrier medium. The self-healing efficiency was evaluated through controlled crack induction followed by curing under favorable environmental conditions. Mechanical properties, including compressive strength, split tensile strength, and flexural strength, were assessed and compared with those of conventional concrete. Crack healing performance was examined using visual inspection, water permeability tests, and microscopic analysis. The results indicate a notable enhancement in crack closure and a significant reduction in water permeability in bacterial concrete specimens. Additionally, bacterial incorporation contributed to improved long-term mechanical strength due to bio-mineralization and pore refinement. The study demonstrates that bacterial self-healing concrete offers an effective, eco-friendly solution for improving structural durability, reducing maintenance requirements, and extending the service life of concrete infrastructure. The findings highlight the potential of bio-based materials as a promising advancement toward sustainable construction practices.

Design And Analysis Of Steel Structures Using Tekla And STAAD

Authors: Satuluri Jhansi Rani, G. Nagalakshimi

Abstract: The design and analysis of steel structures require accuracy, efficiency, and strict compliance with design codes to ensure safety and economy. Modern structural engineering practice increasingly relies on advanced software tools to handle complex geometries, load combinations, and detailing requirements. This study presents a comprehensive approach to the structural analysis and design of steel structures using STAAD for analysis and Tekla Structures for detailed modeling and fabrication-level detailing. Structural analysis is performed in STAAD to evaluate member forces, displacements, and stability under various loading conditions, including dead, live, wind, and seismic loads as per relevant design standards. The designed member sections are then transferred to Tekla Structures to generate precise three-dimensional models, connection details, shop drawings, and material take-offs. The integration of analysis and detailing platforms enhances design accuracy, minimizes errors, and improves constructability. The study demonstrates how coordinated use of STAAD and Tekla leads to optimized structural performance, reduced design conflicts, and efficient fabrication planning. The results confirm that software-based integrated workflows significantly improve the reliability and productivity of steel structure projects.

Impact Of Component Modifications On Vehicle’s Understeer Gradient Reduction

Authors: Raghunath Parate, Priyanka Ranalkar, Gayatri Medge

Abstract: This paper investigates the effects of various suspension and steering system modifications on reducing understeer gradient in a front-wheel-drive (FWD) passenger car. The modifications tested include changes to tyre pressure, rear spring stiffness, steering geometry, and lower control arm (LCA) bushings. Dynamic handling trials on the vehicle were conducted to assess the impact of these modifications. Results show that certain modifications led to substantial reductions in understeer and can bring the undesired high understeer within required limits, while others had no measurable effect. The overall aim was to optimise vehicle handling, improve cornering performance, and enhance driver control.

Isolation And Characterization Of Probiotic Bacteria From Fermented Foods

Authors: Dr. Thara N K

Abstract: The present paper provides a thorough analysis of the methodologies and results associated with the isolation and characterization of probiotic bacteria from traditionally fermented foods. This research, through the thorough analysis of recent studies from 2021 to 2026, investigates the screening, identification, and characterization of lactic acid bacteria (LAB) from various fermented food products, such as cereals, palm nectar, and non-dairy Indian fermented foods. This study proposes a novel framework for the characterization of probiotic bacteria, known as the Multi-Phase Probiotic Characterization Framework (MPPCF), by employing traditional culture-dependent methodologies and molecular characterization. Analysis of the results indicates that fermented foods contain various LAB species, i.e., Lactiplantibacillus plantarum, Lactococcus lactis, Latilactobacillus curvatus, and Leuconostoc mesenteroides, exhibiting strong probiotic potential. The key functional attributes include acid tolerance (survivability of 17-21% at pH 2), bile resistance (28-35% survivability at 0.3% bile salts), antibacterial activity against human pathogens such as P. aeruginosa and S. aureus, and antioxidant activity with scavenging ability higher than 80% for DPPH. The comparative evaluation of the five dimensions of analysis—source diversity, identification techniques, probiotic standards, functional attributes, and safety assessment—establishes the fact that traditional fermented foods are an underutilized reservoir of new probiotic microorganisms with great promise in the development of functional foods and therapeutic formulations.

DOI: https://doi.org/10.5281/zenodo.19109087

Gamification Strategies To Improve Motivation And Retention In Secondary Education

Authors: Dr. Prakash H S

Abstract: The paper offers an extensive analysis of gamification strategies for increasing students’ engagement and academic retention in secondary education. The study undertakes a thorough analysis of existing literature from 2021 to 2026 to examine the impact of specific gamification elements such as points, badges, leaderboards, collaborative gamification, and storytelling integration on students’ engagement and academic retention. The study offers a Dual-Pathway Gamification Framework (DPGF) that differentiates between achievement and immersion gamification. The study is based on the integration of Self-Determination Theory, Cognitive Load Theory, and Perceived Value Theory. The study undertook an extensive analysis of recent experiments from 2021 to 2026 and found that points and instant feedback show the highest correlation with course completion and engagement. Collaborative gamification elements show a high impact on both academic engagement and enjoyment. The study also found that badges and leaderboards show minimal correlation. Moreover, the meta-analytical findings suggest that gamification significantly impacts the perceptions of autonomy support and relatedness among students. However, the impact on the perception of competence is low. The comparison of gamification along the analytical dimensions of motivation type affected, engagement mechanism, retention impact, optimal implementation context, possible drawbacks, and support base indicates the importance of the integration of gamification components with the characteristics of the learners and the context, including family communication style. The findings suggest the importance of gamification designs that promote progress, autonomy, and relatedness.

DOI: https://doi.org/10.5281/zenodo.19109295

MohallaHub: A Hyperlocal Digital Platform For Community – Centric Interaction

Authors: Geena Varghese

Abstract: In the modern digital era, social networking platforms have transformed communication by connecting individuals globally. However, these platforms often lack local relevance and fail to strengthen neighbourhood-level interactions. This paper introduces MohallaHub, a hyper local digital platform designed to foster community-centric interaction within verified geographical boundaries. The system ensures secure communication, promotes trust-based commerce, and enhances local collaboration. By integrating features such as community discussions, marketplace transactions, service listings, and auction systems, MohallaHub creates a structured digital ecosystem tailored for neighbourhood engagement. The platform utilizes a MERN-based architecture to ensure scalability, performance, and data integrity.

Spatial And Vertical Variability Of Petroleum Hydrocarbon Contamination In Depth-Stratified Soils Following A Sabotage-Induced Pipeline Spill: A Case Study From The Niger Delta, Rivers State, Nigeria

Authors: Ernest Ikotiko, E, Osayande, A.D

Abstract: Accurate understanding of the spatial and vertical distribution of petroleum hydrocarbons in post-spill soils is very important to develop effective intervention strategies, assess the processes of natural attenuation of spill impacts, and conduct risk assessment according to regulatory requirements. This study examined the spatial variability and vertical distribution of total hydrocarbon content (THC) and associated physicochemical properties in soils impacted by a sabotage incident that ruptured a pipeline at Obele-Ibaa, Emohua Local Government Area, Rivers State, Nigeria, on April 14, 2015. Soils were sampled approximately 1 year post-spill from three paired topsoil and subsoil locations at depths of 0-15 cm and 15-30 cm, respectively, and one uncontaminated control site. The total hydrocarbon content was determined using n-hexane as a solvent and ultraviolet spectrophotometry as the analytical method. Spatial variability was characterized using contamination factors (CF), coefficient of variation (CV), and vertical distribution ratios (VDR = subsoil THC/topsoil THC). The THC values ranged from 1,302.2 to 1,860.7 mg/kg with a mean of 1,495.6 mg/kg and a coefficient of variation of 17.3 % in the topsoil samples, indicating moderate uniformity of surface contamination. In contrast, the THC values ranged from 769.6 to 4,443.6 mg/kg with a mean of 2,021.2 mg/kg and a coefficient of variation of 84.8 % in the subsoil samples, indicating high spatial variability with vertical distribution ratios ranging from 0.59 to 2.39. This study has shown that there is a two-tier contamination architecture with a uniform surface layer formed by initial spill impacts and a spatially heterogeneous subsoil layer that is governed by preferential flow through macropores and structural discontinuities in the soils. Position P3 was identified as a sub-surface hotspot with high THC values of 4,443.6 mg/kg and a high vertical distribution ratio of 2.39, indicating preferential vertical flow through a high permeability path. All the contamination factors were above the Department of Petroleum Resources (DPR) threshold of 50 mg/kg by 15 to 2,020-fold. This study has shown that spatially uniform intervention strategies are inappropriate for spill impacts of this class and that sub-surface hotspots need to be targeted during cleanup operations. A spatial remediation prioritization framework based on Tier I, Tier II, and Tier III is proposed based on contamination factor values and vertical distribution ratio values.

Mindcare360 – A Mental Fatigue Detector and Personality Based Relief Planner

Authors: Dr.K. Geetha, Boomija.C. B, Dheepa Laksmi.A

Abstract: Within the contemporary digital landscape, individuals often encounter mental fatigue stemming from extended screen time, disrupted sleep cycles, and occupational stressors. The early identification of mental fatigue is crucial for sustaining productivity, emotional equilibrium, and overall health. Conventional wellness applications typically offer broad suggestions that do not account for individual variations in personality and lifestyle. This constraint diminishes the efficacy of these systems in mitigating mental fatigue[1]. This study introduces MindCare 360, an intelligent wellness framework that employs machine learning and explainable artificial intelligence (XAI) to identify mental fatigue and formulate personalized strategies for relief planner. The proposed system employs dual Random Forest classifiers to predict fatigue levels and personality types based on behavioral and lifestyle attributes such as sleep duration, stress levels, physical activity, social interaction, and screen time. Customized daily wellness strategies, segmented into morning, afternoon, evening, and night routines, are generated by integrating the outputs of these models through a rule-based decision-making process. Predictions, confusion matrices, and analyses of feature impact are presented via a Streamlit dashboard, designed with user experience as a priority. Furthermore, the Explainable AI component enhances transparency and builds user trust by elucidating the influence of various lifestyle factors like sleep hours ,sleep quality, and stress score based on their reactions to various situvations on fatigue predictions. Experimental assessments indicate that the proposed system accurately predicts fatigue levels while offering actionable and personalized relief suggestions. The MindCare 360 framework underscores the potential of integrating predictive analytics with personalized wellness planning to aid proactive mental health management.

DOI: https://doi.org/10.5281/zenodo.19111101

Dimensionality-Reduced Frame Work for Predictive Credit Scoring

Authors: Dr. G.Ramasubba Reddy, Dudekula Sreya, Bogireddy venkata Sravani, Guvvala Lasya, Lomada Sanjay Kumar Reddy

Abstract: The growing complexity and scale of financial and behavioral datasets have posed significant challenges to traditional credit scoring methods, as conventional models often fail to capture temporal dependencies and nonlinear feature interactions. To address this, a Hybrid Long Short- Term Memory (LSTM) network was designed, incorporating temporal features for accurate credit scoring prediction. The model integrates a hybrid loss function combining binary cross entropy for classification tasks and optimization techniques such as Min-Max normalization, Synthetic Minority Oversampling Technique (SMOTE) for imbalanced data handling, Recursive Feature Elimination (RFE) for feature selection, and Principal Component Analysis (PCA) for dimensionality reduction. Performance evaluation was conducted on benchmark datasets including the Credit Risk dataset, which provides both structured and unstructured financial data. Comparative analysis against baseline models such as Random Forest and XGBoost demonstrated the effectiveness of the approach. Furthermore, a self- attention mechanism was incorporated into the LSTM framework to enhance contextual learning by emphasizing critical input features, leading to improved predictive accuracy. Experimental results indicate that the Hybrid LSTM with self-attention achieved superior performance with 89.87% accuracy, outperforming existing machine learning and deep learning techniques in credit score prediction.

DOI: https://doi.org/10.5281/zenodo.19126434

Injury patterns in sports-related knee and shoulder injuries in Wrestlers from Haryana Practicing in Government Academies

Authors: Dr. Anjli

Abstract: Wrestling is an Olympic sport for both men and women. It is so ancient since 708 B. C. It is a combat sports that results in various types of injuries that occur in sports. Wrestling and football, which involve frequent player-to-player contact, have a high risk of shoulder injuries. Furthermore, wrestling involves constant contact with the playing surface, exacerbating the condition According to research from the Center for damage and Policy, football and wrestling are the two sports with the highest risk of major damage to athletes. Aim of this study is to know about injury patterns in sports-related knee and shoulder injuries in wrestlers from Haryana practicing in Government Academies. Data has been collected through off line structured questionnaire of wrestlers and coaches of Government Academies of Haryana State. Sample size for the collection of the data has been collected from the total 551 respondents (wrestlers) on the basis of exclusion and inclusion criteria. In our study population, 40 out of 551 players (7.25%) who experienced shoulder injuries and 77 out of 551 (13.97%) who experienced Knee injuries. In our study, 7.25% of participants reported shoulder injuries.

DOI: https://doi.org/10.5281/zenodo.19099736

Human-Following Robot Using Arduino Uno

Authors: Dr. Suchita Walke, Sojal Patil, Yash Gadge, Sudarshan Shedge

Abstract: Nowadays, Technology Advances So Rapidly That Robots Start to Emerge Not Just in Factories, But in Our Daily Lives Helping Us With Big And Small Things. For This Project, We Were Excited to Build Something Helpful and Entertaining: A Human-Following Robot. It Is a Small, Moving Robot Built Out of Low-Budget and Beginner- Friendly Materials Like an Arduino Uno, Ultrasonic Sensors, And A Motor Driver. The Idea Is Simple to Grasp the Robot Follows A Person By Detecting Their Movement And Staying A Safe Distance Behind. This Is How It Works: The Robot Has Ultrasonic Sensors That Send Out Sound Waves. When The Waves Reflect Off Something, The Robot Calculates How Far Away That Something Is. With Two Or More Sensors, It Can Even Calculate The Direction The Person Is Heading. The Arduino Does That Entire Math and Tells the Motors How to Go So The Robot Can Coast Along Smoothly Without Running Into Things. This Robot’s Design Is Versatile and Has Numerous Applications In Home Life. It Can Be an Intelligent Shopping Cart That Accompanies You In A Supermarket, A Home Robot That Assists The Elderly, Or A Warehouse Robot That Transports Items From One Point To Another. And as The Design Is Low-Cost and Minimalistic, It Is Perfect For Hobbyists Who Want To Play With Robotics, Sensors, And Embedded Systems. Once The First Version Is Up and Running, There’s Plenty Of Scope For Enhancement Such As Incorporating A Camera For Video Tracking, Incorporating Ai To Make It Intelligent, Or Networking It For Net-Based Monitoring. For Hobbyists and Students, It’s An Excellent Hands-On Exercise To Learn How Software And Hardware Interface, And A Precursor To Building Sophisticated and Intelligent Robots In The Future.

DOI: https://doi.org/10.5281/zenodo.19132010

CropGuard: A Mobile-Based Plant Health Detection System Using Convolutional Neural Networks

Authors: Prof. Maske P. P, Kasturi Kumbhar, Utkarsha Chipade

Abstract: Agriculture plays a vital role in ensuring food security and supporting the global economy. However, plant diseases remain one of the major challenges that significantly affect crop productivity and quality. Early detection of plant diseases is essential to prevent large-scale damage and to help farmers take timely preventive actions. Traditional methods of disease detection rely on manual inspection by agricultural experts, which can be time- consuming, expensive, and sometimes inaccurate. With the advancement of artificial intelligence and deep learning technologies, automated plant disease detection systems have become increasingly effective. This research presents CropGuard, a mobile-based plant health detection system that uses Convolutional Neural Networks (CNN) to identify plant diseases from leaf images. The proposed system allows farmers to capture images of plant leaves using a smartphone camera. The captured image is processed using image preprocessing techniques such as resizing and normalization before being analysed by a trained CNN model. The model extracts important visual features from the leaf image and classifies it as healthy or diseased based on patterns learned during the training phase. The system then displays the predicted disease information through the mobile application interface, allowing farmers to quickly understand the health condition of their crops. The dataset used for training consists of labelled images of healthy and diseased plant leaves, enabling the CNN model to achieve reliable performance. The proposed system demonstrates how artificial intelligence and mobile technology can be combined to provide a fast, accurate, and user-friendly solution for plant disease detection and smart agricultural practices.

Integrating Solar Energy On Water Bodies For Sustainable Renewable Energy Development

Authors: Purvi R, R.M.S Keerthi, Rachana, Vikas Naregal, Vishal S, Hesham F, Dr. Prithi Madhavan

Abstract: Floating solar farms (FSFs) are an innovative way of generating renewable energy; this approach addresses certain major shortcomings of traditional land- based solar power plants. FSFs offer the Two fold benefits of both renewable energy production and land preservation through placing panels of photovoltaic cells in water bodies, including lakes, reservoirs, and coastal waters. Since natural cooling of water surfaces is realized, it also contributes to increasing panel efficiency and reducing evaporation of water. This study discusses the economic and environmental advantages of FSFs such as higher energy and potential water savings in regions that are experiencing shortages. The major challenges of high costs of installation, recurrent maintenance and potential ecological impact on aquatic life are also harshly analyzed. The paper provides a comparative overview on the world experiences and perceptions on the potential of FSF technology in the future based on the case studies in China, India, and Netherlands. The data are supported with graphs and figures that indicate cost- performance, energy efficiency outcomes, and acceptance trends of FSF. Ultimately, in this paper the author highlights the relevance of floating solar farms in the realization of global sustainability goals and addressing the increasing demand of renewable energy. Keywords- FLOating Solar Farms, renewable energy, sustainability, Photovoltaic technology, water conservation.

Gradmate Ai

Authors: Akash Shiboy, Amritha S, Anjali Unnikrishnan, Elvin Manoj George, Dr. Rani Saritha R

Abstract: Gradmate AI is a web-based academic and placement management system that integrates artificial intelligence to improve learning efficiency and streamline institutional processes. Developed using Python Flask and Firebase, the system ensures secure, scalable, and real-time data management. It incorporates Gemini AI to provide features such as chatbot assistance, content summarization, and quiz generation, enhancing student engagement and reducing manual effort. The platform supports both students and placement officers by enabling study planning, resume management, and placement tracking within a unified environment.

Thyroid Disease Prediction

Authors: Albert S. Joseph, Annie David, Jude Joby Joseph, Dr. Rani Saritha R

Abstract: Thyroid disorders represent a significant global health challenge, affecting approximately 5% of the population and necessitating precise screening to prevent risks like cardiac arrhythmias and metabolic imbalances. This research addresses the limitations of manual diagnostics—often subjective and prone to error—by introducing an Automated Thyroid Diagnostic Assistant. Leveraging the XGBoost machine learning algorithm, the system classifies patient status into four distinct categories: Normal, Primary Hypothyroid, Compensated Hypothyroid, and Hyperthyroid. The model was developed using the UCI Thyroid Disease dataset, utilizing 18 critical features, including demographics, medical history, and hormone levels ($TSH$, $T3$, and $T4$). To ensure clinical utility, the system is deployed via a Flask-based web interface, providing medical professionals with near real-time predictions and confidence scores.

Structural Design and Performance Evaluation of a 3D Printer

Authors: Balamurugan J, Logeshwaran D, Mohan E, Suresh G

Abstract: Structural design and performance evaluation of a 3D printer, focusing on improving stability, dimensional accuracy, and overall print quality. In many low-cost fused deposition modelling (FDM) printers, structural limitations such as weak frames and improper alignment lead to vibrations and inconsistent results. To address these issues, this study emphasizes the development of a rigid and well-aligned mechanical structure that minimizes unwanted motion during operation. The printer was assembled carefully with attention to frame rigidity, motion system alignment, and extrusion stability. Experimental evaluation was carried out using dimensional accuracy testing, print speed versus quality analysis, and repeatability assessment.

Verifying Digital Academic Credential Using Blockchain

Authors: Ms. Pradnya Patil, Omkar Karade, Soham Vare

Abstract: In today’s increasingly digital and global environment, traditional methods of issuing and verifying certificates are prone to fraud, loss, delays. This solution introduces a blockchain-based system for storing and protecting academic credentials, where educational institutions issue digital certificates directly onto a blockchain network based webapp, Blockchain network mean’s Polygon Amoy. In which, the user will be able to have all the control related to it’s credentials and not the institution, as sometimes personal information of users can get leaked by institutions which is a breach of smart contract, so this webapp will ensure full control to the users. Through the use of smart contracts, the system automates the verification process, which will be directly issued to the intended user and also will adapt paperless system which will be environment friendly. This approach not only reduces the dependency on educational institutions it will also help to save paper which is the major factor for such credentials/certificates.

IoT – Based Real Time Patient Health Monitoring System Using Raspberry Pi And Cloud Integration

Authors: Dhanya sri S, Barath kanna G, Dr. R. Karthik

Abstract: Continuous monitoring of patient health parameters is a critical requirement in modern healthcare systems, particularly for elderly patients, individuals with chronic conditions, and post-operative care scenarios. Traditional hospital-based monitoring systems are expensive, stationary, and impractical for home-based care environments. This project proposes an IoT-Based Real-Time Patient Health Monitoring System that uses a Raspberry Pi microcomputer interfaced with multiple biometric sensors to continuously collect and transmit vital health parameters including body temperature, heart rate, blood oxygen saturation level, and blood pressure readings. The collected data is transmitted wirelessly to a cloud platform where it is stored, processed, and made accessible through a web-based dashboard. Automated alert mechanisms notify healthcare providers and designated caregivers when recorded values fall outside predefined safe thresholds. The system is designed to be compact, affordable, and easy to operate without requiring technical expertise from patients or caregivers. Testing results confirm that the system achieves high sensor accuracy and reliable data transmission under real-world operating conditions. This research demonstrates how Internet of Things technology can be integrated with cloud computing to deliver an accessible and effective remote patient monitoring solution that reduces healthcare costs and improves patient outcomes.

DOI: https://doi.org/10.5281/zenodo.19183571

Forensic Anatomy in Cyber Crime: Bridging Biological Forensics and Digital Investigations for Deepfake Authentication

Authors: Payal Panda, Sai Aditya, Krishna Sravya Chirla

Abstract: With the rise of Generative AI models that are achieving near-pixel perfection, the traditional digital forensic methods, which rely on noise analysis and metadata, are becom- ing increasingly inefficient. Cyber-criminals and scammers are now using Generative AI models to generate synthetic media, which can be used for identity theft, spreading false information, and financial fraud. This advancement in the digital world requires us to shift from digital-only detection to anatomical or biological validation. This research introduces a safety frame- work that bridges the gap between clinical forensics and digital investigations. Real human faces exhibit muscle movements and heart pulses, unlike deepfake faces, which enable us to use the human body as a biological watermark. We use remote Photoplethysmography (rPPG) to observe and extract the pulse signals from the forehead and the cheeks, while a clinical audit led by anatomical experts evaluates musculoskeletal synergy to identify the biological impossibilities present in the synthetic media. This safety framework proposes a physiology-informed forensic framework for legal and investigative usage, providing a chance to combat cyber-criminals and scammers.

DOI: https://doi.org/10.5281/zenodo.19184288

Quantum Computing For Combinatorial Optimization: Algorithms, Complexity Analysis, And Real-World Applications.

Authors: G. Swapna, P. Sunil

Abstract: Combinatorial optimization problems play a vital role in computer science and engineering, with applications in logistics, network design, finance, and resource allocation. However, many of these problems are NP-hard, making them computationally expensive to solve using classical optimization techniques, especially for large-scale instances. In recent years, quantum computing has emerged as a promising paradigm capable of addressing these limitations by leveraging quantum mechanical principles such as superposition and entanglement. This study investigates the application of quantum computing for solving combinatorial optimization problems, with a focus on algorithm design, complexity analysis, and real-world applicability. The research primarily employs the Quantum Approximate Optimization Algorithm (QAOA) and Grover’s search algorithm to solve representative problems such as Max-Cut and the Knapsack problem. The performance of these algorithms is evaluated using key metrics including accuracy, execution time, and approximation ratio, and is compared with classical optimization techniques. The results demonstrate that QAOA is capable of generating near-optimal solutions, achieving accuracy levels of up to 93% for small problem instances. However, the study also highlights the challenges associated with current quantum systems, including noise, limited qubit availability, and increased execution time. Comparative analysis reveals that while classical methods remain more efficient for small-scale problems, quantum algorithms show significant potential for scalability and handling complex optimization tasks. The findings suggest that hybrid quantum-classical approaches offer a practical solution in the current Noisy Intermediate-Scale Quantum (NISQ) era. The study concludes that although quantum computing is still evolving, it holds strong potential for transforming combinatorial optimization in the future.

An Intelligent System For Early Leak Prediction And Risk Localization In Urban Water Pipelines

Authors: K. Madhumitha, Nandhakumar G, S Prasanna, A Pragatishvar

Abstract: Urban water distribution systems are essential for supplying clean and safe water to residential, commercial, and industrial sectors. However, aging infrastructure, pipe corrosion, pressure fluctuations, and environmental factors often lead to pipeline leaks that result in significant water loss, economic damage, and potential public health risks. Early detection of such leaks is therefore critical for efficient water management and infrastructure maintenance. This paper presents an intelligent system for early leak prediction and risk localization in urban water pipelines using advanced data analysis and sensor-based monitoring techniques. The proposed system collects real-time data from pressure sensors, flow meters, and acoustic sensors installed throughout the pipeline network. Machine learning algorithms analyze the collected data to identify abnormal patterns that indicate potential leakage. The system also employs predictive models to estimate the likelihood of future leaks and determine high-risk pipeline segments. By integrating data analytics, sensor networks, and intelligent decision-making methods, the proposed approach enables early warning and accurate localization of leaks. This helps water management authorities reduce water loss, improve maintenance planning, and enhance the reliability and sustainability of urban water distribution systems.

DOI: https://doi.org/10.5281/zenodo.19184732

Plot Point : The Real Estate Hub

Authors: Abel John Jacob, Abhimol Manoj, Devadathan P.R, Liya Sara Joseph, Dr. Rani Saritha R

Abstract: The real estate sector still depends heavily on manual processes for maintaining land records, managing customer enquiries, verifying documents, and confirming plot bookings. These traditional approaches result in delays, inconsistent data management, limited transparency, and a high risk of human error. Plot Point is a web-based, centralized land and plot management system developed to address these challenges by automating the complete workflow of plot booking and real-estate documentation. The proposed system integrates modules for staff authentication, customer enquiry handling, real-time plot availability tracking, digital document submission and verification, automated agreement generation, and booking confirmation. Through its structured workflow and secure data management architecture, Plot Point minimizes manual workload, enhances accuracy, and ensures traceability throughout the booking lifecycle. The system also supports faster decision-making by providing staff with consolidated dashboards and well-organized records. This digital transformation improves operational efficiency, reduces administrative overhead, and enhances customer satisfaction by enabling transparent, error- free, and time-efficient plot booking operations.

DOI:

Peristaltic Transport of a Newtonian Fluid in an Asymmetric Channel with Wall Slip: Influence of Waveform Shapes

Authors: K. Rajanikanth

Abstract: This study investigates the effects of wall slip and various waveforms on the peristaltic transport of a Newtonian fluid in a two-dimensional asymmetric channel. Channel asymmetry is generated by imposing peristaltic wave trains of different amplitudes and phase differences on the upper and lower channel walls. The mathematical formulation is developed under the assumptions of long wavelength and low Reynolds number. Exact analytical solutions for the stream function, axial velocity, and pressure gradient are obtained. Numerical computations are performed to analyze pumping characteristics, frictional forces, trapping, and reflux phenomena. Results show that increasing the permeability parameter reduces pumping against pressure rise, axial velocity, pressure gradient, trapped bolus size, and reflux layer thickness. For sufficiently large permeability, symmetry of the trapped bolus is lost. Under certain conditions, closed streamlines form, resulting in trapped boluses moving with the wave speed. A comparative analysis of four waveforms—triangular, sinusoidal, trapezoidal, and square—indicates that the square waveform yields the highest volumetric flux.

DOI: https://doi.org/10.5281/zenodo.19202495

The Electric Vehicles Revolution: The World Moving toward a Cleaner and Sustainable Transportation

Authors: MD. Mobarak Hossain, Kashinath SA, Surya Pratabp Singh

Abstract: The global movement of people and goods is deteriorating the quality of air by introducing increased carbon to the atmosphere, and we are still addicted to oil. The use of gasoline-powered cars contributes significantly to the planetary warming. Due to this mess, battery-run cars are making an entry as a cleaner mode of transportation as opposed to the usual models. These cars save on money in the long-term by reducing pollution using an electric motor powered by batteries. One of the changes that have been observed in the past decade across the globe is a result of smarter technology, lower cost of storage units, and assistance in the form of national policies. Governments support electric transport not only in big cities but also in new areas to ensure cleaner air but also in the financial balance in the long term. In retrospect, the history of the development of electric cars is important than it initially appears. Growth did not occur in a single area – technology was being advanced, policies were facilitating it, people began to make different decisions. The changes in habits across the countries expose trends that would not be so apparent. These machines are not just moving around – they are making emissions and they are doing it with a low profile. Their significance infiltrates urban areas, streets, lives. In the long run, they are useful where old models fail. Until the time it does not come, change comes slowly. It is possible that the future of travel is greatly due to what started as small substitutes.

DOI:

Rubyfi: The Wi-Fi Red Teaming Toolkit

Authors: Abhinav Ugale, Omkar Shinde, Manmeet Singh, Ms.Pradnya Patil

Abstract: Wi-Fi is now used everywhere, from our homes to public locations, and it connects lakhs of people. However, this ease comes with a major drawback. Most networks contain security flaws that attackers can easily exploit. These weaknesses can cause major problems such as data theft and invasion of privacy. To properly test for these dangers, one has to use many different and complicated command-line tools, which is not practical for everyone. To solve this, we have developed RubyFi, a complete Wi-Fi security testing toolkit. Our system brings all the necessary tools like aircrack-ng, reaver, and hashcat into a single, easy-to-use platform with both a graphical and command-line interface. RubyFi allows a user to perform a full range of tests, from capturing WPA/WPA2 handshakes for password cracking to launching Deauthentication and Evil Twin attacks. Importantly, we have also included modules to test for the latest WPA3 vulnerabilities, such as those found in the Dragonblood research, making it relevant for modern networks. Finally, RubyFi is an instructional tool that aims to make cybersecurity realistic. It makes ethical hacking easy, providing both students and security pros a straightforward way to find and understand Wi-Fi weaknesses. The main goal is to help people spot these security holes and learn how to fix them, leading to safer wireless networks for everyone.

CopilotX: A Next Generation AI Assistant for Human-Computer Interaction

Authors: Shahid Ali, Ashutosh Rautela, Mr. Sachin Misra

Abstract: Artificial Intelligence (AI) has emerged as one of the most revolutionary technologies in contemporary computing, particularly in the fields of education and software development. This review paper examines the implementation of Artificial Intelligence in the significant project “CopilotX – Your AI Future,” which is an AI- driven assistant intended to aid students and developers in coding, debugging, and educational tasks. The paper evaluates existing research on large language models (LLMs), natural language processing (NLP), and AI-based coding assistants. It contrasts current AI copilots with the proposed CopilotX system to discern their advantages and drawbacks. The study underscores challenges such as elevated API costs, insufficient personalization, and the absence of structured academic guidance in existing systems. CopilotX seeks to fill these voids by incorporating conversational AI, intelligent code review, and educational support features into a cohesive platform. The results suggest that AI-driven assistants considerably improve productivity, learning efficiency, and problem-solving skills. Prospective enhancements may encompass adaptive learning models, voice interaction, and domain-specific AI customization. Machine learning, Deep Learning, Neural networks, Natural Language Processing and knowledge base system of the Artificial intelligence Assistant.

DOI: https://doi.org/10.5281/zenodo.19203177

Campus Rideshare

Authors: Abijith Sankar, Amina Rashad, Belvin Thomas Cherian, Elona Elsa Thomas, Dr. Abin T. Abraham

Abstract: College campuses often experience transportation challenges such as overcrowded shuttles, limited affordable commute options, traffic congestion, and unpredictable travel timings. Traditional campus travel systems lack real-time ride availability, secure identity verification, and a structured process for students and faculty to share rides. The Campus RideShare System addresses these issues by providing a secure, web-based platform that connects verified campus members traveling along similar routes. Through institution-based login, users can offer or request rides, match with nearby commuters, and share live location information for a safer and more convenient travel experience. Developed as a modern web application using Firebase and Leaflet for geolocation and mapping, the system incorporates features such as real-time route tracking, intelligent distance-based ride matching, and a rating mechanism to ensure trust and safety. By optimizing campus travel resources, improving coordination, reducing travel costs, and lowering carbon emissions, the proposed solution enhances overall campus mobility while promoting sustainable transportation.Keywords— Ride-sharing, Campus Transportation, Real-time Tracking, Firebase, Geolocation, Mobile Application, Smart Mobility.

Modelling And Performance Analysis Of A Static Electric Vehicle Wireless Charging Pad Under Multiple Coil Misalignment

Authors: Ms. S. Bharathi

Abstract: Static WPT are receiving enormous attention as a future-forward charging solution to enable contactless, automated, and less maintenance powering for EVs. However, the power delivering efficiency and charging stability of inductive coupled static WPT systems strongly deteriorate under spatial misalignment between primary and secondary coils, which is inevitable in real-world parking scenarios. This paper represents the modelling, simulation, and performance evaluation of a inductive coupled static wireless EV charging pad under multiple coil misalignment conditions. A MATLAB/Simulink based high-frequency inverter along with L-C compensation network is developed to emulate realistic EV charging behavior, and the effect of lateral displacement on magnetic coupling, power flow, and State-of-Charge (SoC) trajectory is analyzed. Misalignment scenarios of 0%, 20%, and 40% are emulated by varying the coupling coefficient (k) from unity to reduced values, and the corresponding impact on induced voltage, input/output power, and energy efficiency is investigated. Results depict progressive efficiency degradationfrom ~92% at perfect alignment to ~65% under severe displacement, and it is accompanied by remarkable reduction in charging current and dynamic voltage ripple. This study confirms the sensitivity of WPT systems to coil offset and brings into perspective the necessity for adaptive compensation, real- time control, and misalignment tolerant coil topology. The findings serve as a design benchmark and provide valuable insights for integrating intelligent tuning algorithms and optimization frameworks to ensure reliable, high-efficiency wireless charging in future EV infrastructures and smart transportation ecosystems.

ARCH: An LLM-Driven Autonomous Self-Healing Framework For Cloud Operations Using Retrieval-Augmented Generation

Authors: Ripusoodan Sharma, Dr. Kirti Jain

Abstract: The increasing complexity of cloud-native environments has made fault detection and recovery a critical challenge for modern IT operations. Traditional AIOps solutions rely heavily on static rules or data-driven models, which often lack adaptability in dynamic and unpredictable scenarios. To address these limitations, this paper proposes the ARCH (Autonomous Reasoning and Contextual Healing) framework, an intelligent self-healing system that integrates Large Language Models (LLMs) with Retrieval-Augmented Generation (RAG) for autonomous cloud operations. The proposed framework follows a layered architecture consisting of perception, cognition, knowledge, and action components, enabling continuous monitoring, contextual reasoning, and automated remediation. By leveraging advanced reasoning strategies such as chain-of-thought and action-oriented decision-making, the system dynamically analyzes telemetry data and executes corrective actions with minimal human intervention [5], [7]. Furthermore, the incorporation of RAG enhances the system’s ability to utilize historical incident data, thereby improving decision accuracy and contextual awareness [4]. The effectiveness of the ARCH framework is evaluated using key performance metrics, including Mean Time to Repair (MTTR), Autonomous Success Rate (ASR), and overall system efficiency. Experimental results demonstrate that the proposed approach significantly improves fault resolution performance, achieving up to 82% reduction in MTTR and 89.5% autonomous success rate compared to conventional methods. These findings highlight the potential of LLM-driven architectures in enabling scalable and intelligent self-healing cloud systems.

DOI: https://doi.org/10.5281/zenodo.19204589

Diabetes Prediction System Using Machine Learning And Web-Based Interactive Tool

Authors: Dr. D. Siva Sankara Reddy, R.S. Safiya, M. Naga Venkat, P. Anil, Jagadeeswar Reddy

Abstract: The prevalence of diabetes is rising globally, making early detection crucial for effective management and prevention of complications. This project aims to develop an end-to-end machine learning application to predict the likelihood of diabetes in patients based on diagnostic measurements. Using the Pima Indians Diabetes Database, we employed a Random Forest Classifier to build a predictive model. The model is deployed as a web application using Flask, allowing users to input medical details (e.g., Glucose, BMI, Age) and receive real-time predictions. The project includes data gathering, descriptive analysis, data visualizations, data preprocessing, model building, and model deployment on Heroku.

DOI: https://doi.org/10.5281/zenodo.19204840

The Problem of Overweight in Children Aged 8-10 Years

Authors: Dean Qefalia, Albjola Maloku

Abstract: Static WPT are receiving enormous attention as a future-forward charging solution to enable contactless, automated, and less maintenance powering for EVs. However, the power delivering efficiency and charging stability of inductive coupled static WPT systems strongly deteriorate under spatial misalignment between primary and secondary coils, which is inevitable in real-world parking scenarios. This paper represents the modelling, simulation, and performance evaluation of a inductive coupled static wireless EV charging pad under multiple coil misalignment conditions. A MATLAB/Simulink based high-frequency inverter along with L-C compensation network is developed to emulate realistic EV charging behavior, and the effect of lateral displacement on magnetic coupling, power flow, and State-of-Charge (SoC) trajectory is analyzed. Misalignment scenarios of 0%, 20%, and 40% are emulated by varying the coupling coefficient (k) from unity to reduced values, and the corresponding impact on induced voltage, input/output power, and energy efficiency is investigated. Results depict progressive efficiency degradationfrom ~92% at perfect alignment to ~65% under severe displacement, and it is accompanied by remarkable reduction in charging current and dynamic voltage ripple. This study confirms the sensitivity of WPT systems to coil offset and brings into perspective the necessity for adaptive compensation, real- time control, and misalignment tolerant coil topology. The findings serve as a design benchmark and provide valuable insights for integrating intelligent tuning algorithms and optimization frameworks to ensure reliable, high-efficiency wireless charging in future EV infrastructures and smart transportation ecosystems.

DOI: https://doi.org/10.5281/zenodo.19208100

IMPACT OF ARTIFICIAL INTELLIGENCE IN RECRUITMENT AND SELECTION PROCESS

Authors: Dr. R. INDRA, MS P. BRINDHA

Abstract: The rapid growth of Artificial Intelligence (AI) is transforming recruitment and selection in Human Resource Management by making hiring faster, more efficient, and data-driven. Traditional methods that relied on manual resume screening and interviews were time-consuming and often biased, whereas AI tools such as machine learning, chatbots, and natural language processing now automate candidate sourcing, screening, and evaluation. Platforms like LinkedIn, HireVue, and Pymetrics help organizations reduce hiring time and improve talent matching. While AI promotes diversity and efficiency, it also raises concerns about bias, privacy, and reduced human interaction, especially in diverse contexts like India. Understanding both the benefits and limitations of AI is essential for ensuring fair and effective hiring practices.

DOI:

Mathematics As A Subject: Understanding Math Anxiety In Comparison With Other Subjects In Sirohi, Rajasthan

Authors: Vaibhav Tambe, Vishnu Prakash Kareepadath

Abstract: This Research investigates students’ perspectives on mathematics compared to other subjects, focusing on the plurality and underlying factors of math anxiety in the context of primary and secondary education in Sirohi, Rajasthan. Acknowledge the crucial role of mathematics in the STEM field and the documented phenomenon of math anxiety. This qualitative study aims to understand how students perceive mathematics in terms of interest, difficulty, and anxiety levels when compared with other subjects. Drawings from the National Curriculum Framework 2023 (NCF2023), Which Identifies societal views and teaching methods as key factors contributing to math anxiety. This research used semi-structured interviews with over 30 primary and secondary school students ( grades 5th – 10th, aged 11 to 16) from rural Hindi-medium government schools and English-medium government schools in Sirohi. The collected data was analyzed using thematic analysis. Key findings reveal significant influence by conceptual clarity, effective teaching methods, utilizing familiar examples, and parental support. Furthermore, challenges in teaching approaches subject-specific difficulty (notably areas like algebra, Trigonometry, Sanskrit vocabulary, and Social Science concepts) and subject-related anxiety were reported for mathematics and subjects like Social Science and Sanskrit. Notably, societal expectations position mathematics as the key to success and employment, along with parental influence, and most importantly, it shapes students’ attitudes and choices towards mathematics. These insights inform the development of more effective teaching strategies and curriculum design to reduce subject-related anxieties and promote a more positive attitude towards all subjects, including mathematics, within the context of primary and secondary education in Sirohi, Rajasthan.

Aircraft Wing Flatter Phenomenon Using Matlab Simulink Software

Authors: Master Van Huy Khuat, Master Le Phan

Abstract: This paper presents the results of a study investigating the flutter phenomenon of aircraft wings using MATLAB/Simulink. Flutter is a self-excited oscillation of an aircraft wing that occurs when its airspeed exceeds a critical limit. The wing oscillation model was developed within the MATLAB/Simulink environment. The paper details the simulation results and compares them with experimental data found in the aircraft technical manuals from the Air Force Officer School. This method enables the determination of the critical flutter velocity and analyzes how variations in wing parameters affect this threshold. Keywords: Flutter; Air plane wings; Calculate on Matlab.

DOI: https://doi.org/10.5281/zenodo.19217155

Edge–Cloud Collaborative Framework For Real-Time Quality Control In Smart Manufacturing

Authors: Dr. Pankaj Malik, Tohid Khan, Vinayak Pal, Utsav Malviya, Akshita Rathore

Abstract: Smart manufacturing systems require efficient and real-time quality control mechanisms to ensure high product reliability and minimize production losses. Traditional cloud-based inspection systems suffer from high latency, bandwidth limitations, and delayed decision-making, while edge-only solutions are constrained by limited computational resources. To address these challenges, this paper proposes an Edge–Cloud Collaborative Framework for Real-Time Quality Control, integrating edge computing for low-latency defect detection and cloud computing for large-scale analytics and model optimization. In the proposed system, real-time data from industrial sensors and vision systems is processed locally at the edge for immediate defect detection using a lightweight deep learning model, while the cloud layer performs periodic model training, optimization, and global decision support. A dynamic task offloading strategy is implemented to balance computational load between edge and cloud based on latency, bandwidth, and resource availability. The framework is evaluated using the MVTec AD dataset on an edge device (Jetson Nano) integrated with a cloud platform. Experimental results demonstrate that the proposed system achieves an accuracy of 97.5%, precision of 97.1%, recall of 96.8%, and F1-score of 96.9%, with an average inference latency of 45 ms, significantly outperforming traditional cloud-only systems (latency ~150 ms) and edge-only systems (accuracy ~94.5%). Additionally, the collaborative approach reduces bandwidth usage by approximately 40% through local preprocessing at the edge. These results confirm that the proposed edge–cloud collaborative framework provides an effective balance between low latency, high accuracy, and efficient resource utilization, making it highly suitable for real-time quality control in Industry 4.0 smart manufacturing environments.

The Influence Of Cultural Context On The Representation Of Mental Health In World Literature

Authors: Prafulla Dubey, Dr Devangna Pareek

Abstract: This study examines the influence of cultural context on the representation of mental health in world literature. While mental illness is often discussed within medical and psychological frameworks, literary narratives reveal that perceptions of mental health are deeply shaped by cultural beliefs, social norms, religious values, and historical experiences. Through a comparative textual analysis of selected works from Western, African, Asian, and Middle Eastern literary traditions, this paper explores how different cultures construct, interpret, and narrate psychological distress. The study demonstrates that Western literature frequently emphasizes individual consciousness, introspection, and medicalized understandings of mental illness, whereas non-Western literary traditions often frame psychological suffering within collective, spiritual, or socio-political contexts. In many African and Asian narratives, mental health is interconnected with community relationships, family honor, spirituality, and cultural expectations. Similarly, postcolonial literature frequently portrays psychological trauma as intertwined with colonial history, displacement, and identity conflict. By highlighting cross-cultural differences and similarities, this research argues that literature functions not only as a mirror of cultural attitudes toward mental health but also as a powerful tool for challenging stigma and fostering empathy. The findings underscore the importance of culturally sensitive literary analysis and suggest that world literature can contribute meaningfully to global conversations on mental health awareness and destigmatization.

Integrating Explainable Artificial Intelligence In Healthcare: Models, Applications, And Challenges

Authors: Ms. Babita, Dr. Brij Mohan Goel

Abstract: This paper will explore the role of XAI in advancing healthcare systems by examining various explain ability models and techniques, including feature attribution methods, model-agnostic approaches, and interpretable machine learning frameworks. It will highlight key applications of XAI in medical imaging, clinical decision support systems, disease prediction, and personalized medicine, where interpretability will be crucial for ensuring reliability and accountability. Furthermore, the study will discuss emerging challenges such as the trade-off between model accuracy and interpretability, data privacy concerns, lack of standardized evaluation metrics, and integration barriers within real-world clinical settings. Ethical considerations and regulatory requirements will also be analysed to understand the broader implications of deploying XAI in HealthCare. The paper will conclude by emphasizing the need for robust, scalable, and clinically validated XAI solutions that will bridge the gap between complex AI models and human understanding. Future research will focus on developing hybrid models, improving user-centric explanations, and fostering interdisciplinary collaboration to ensure the safe and effective adoption of explainable AI in healthcare.

DOI: https://doi.org/10.5281/zenodo.19218727

Graph Neural Network-Based Fraud Detection In Blockchain Supply Networks

Authors: Dr. Pankaj Malik, Mohammed Hamd, Shyamal Sheorey, Mohd Ayaz Shiekh, Aditya Narayan Sharma

Abstract: Blockchain technology has emerged as a transformative solution for enhancing transparency, traceability, and immutability in supply chain transactions. However, despite its decentralized security architecture, fraudulent activities such as collusive supplier networks, duplicate invoicing, smart contract exploitation, and phantom shipment generation continue to threaten blockchain-enabled supply ecosystems. Traditional machine learning-based fraud detection models analyze transactions independently and fail to capture the complex relational dependencies inherent in multi-tier supply networks. To address this limitation, this paper proposes a Graph Neural Network (GNN)-based fraud detection framework for blockchain supply networks. The proposed approach models blockchain transactions as graph structures, where nodes represent supply chain entities and edges represent transactional interactions. A Graph Convolutional Network (GCN) is employed to learn structural and feature-bas ed representations of transaction networks, enabling the detection of coordinated and network-level fraudulent behaviors. Experimental evaluation was conducted on a simulated blockchain supply chain dataset comprising 50,000 transaction records and 8,200 interconnected entities. The proposed GNN model achieved an accuracy of 95.2%, precision of 94.6%, recall of 93.8%, and F1-score of 94.2%, outperforming traditional classifiers including Logistic Regression (81.4% accuracy), Random Forest (86.7%), and Artificial Neural Networks (89.3%). Furthermore, the proposed framework reduced false positive rates by 27% compared to baseline methods, demonstrating superior capability in identifying collusive fraud patterns. The results confirm that graph-based deep learning significantly enhances fraud detection performance in decentralized supply chain environments. The proposed system provides a scalable and intelligent security layer for blockchain-enabled supply networks.

Durability Assessment Of Concrete Exposed To Marine Environment

Authors: Mekala Sundarji, Sk.Abdulkareem

Abstract: Asphalt pavements are an essential part of the transportation infrastructure that is necessary for Marine environment growth. This research delves into the novel incorporation of recycled plastics into VG30-grade bitumen for use in road building, including polyethylene terep thalate (PET), high- densitypolyethylene (HDPE), and polyvinyl chloride (PVC). Improving infrastructure resilience and reducing environmental effect through the usage of plastic waste are two challenges that this study seeks to solve. Specifically, it seeks to optimize plastic proportions in order to increase pavement durabilityandsustainability. The Marshall Stability (MS) test and other rigorous laboratory experiments provide the basis of an extensive experimental strategy that examines different asphalt compositions. According to the results, the optimal mix for maximum MS is 3.0% waste plastic, made entirely of PET, with 5.5% bitumen. This formulation showcases substantial performance improvements achieved by selective plastic inclusion, and it outperforms standard asphalt mixes in terms of stability by an impressive 73.07%.

Comparative Structural Analysis Of RC Buildings Using STAAD Pro And SAP2000_409

Authors: Shaik Jamal Basha, U.Srinivasarao

Abstract: The project titled: “PLANNING AND STRUCTURAL DETAILDED ESTIMATION COSTING OF G+5 BUILDING AND INTERIOR DESIGN” using “STADD PRO” software. The project gives the overview of planning, structural detailing and cost estimation of g+5 storey building and estimation of interior designing of a building. We know before starting the construction we need to know, how much amount is required for finishing of construction For that we need to go for cost estimation of the building for each and every component work which gives the brief or rough idea of cost the building to finish the construction. By using this estimation we can arrange the money and material according to the requirement and can know the cost of every work done in construction. Planning and structural detailing are the major parts in the building in which construction is progressed and proceeded based on the planning of building and orientation and also the structural detailing of the concrete and steel members. In this project we are dealing with the planning, structural detailing, cost estimation of building and interior designing for g+5 storey building. Planning of building is taken out according to the norms of town planning commission and structural detailing is taken out by following code book required for the designing of member and detailing of member and in the same way cost estimation of building and interior designing are done as per CPM-PERT and morth specification rate of Indian government. Here planning deals with selection of site, orientation of building and placing of rooms as per requirement. Structural detailing deals with the designing of beams columns, slabs and stair case of the building and also giving the structural detailing of steel required to place in concrete member during casting. Also cost estimation deals with the rates of each and every work and the component used in the construction work. In this project we have done every-thing by following the specification and got the accurate results and finished the project successfully.

Ai-Powered Pneumonia Detection Using Chest Ct Images

Authors: Mr. R. Premkumar, Mr.R.Eswara Prakash, Mr. M. Manoj, Mr. K. Janarthanan

Abstract: The aim of this project is to classify CT Scan images of pa- tients with or without pneumonia. More specifically, we trained a Convolutional Neural Network(CNN) of differ- ent parameters with chest CT Scan images of children and the outcome classes are two: ”Pneumonia” or ”Non – pneumo- nia”. The findings follow in the next sections. The primary objective of this work is to develop an efficient and reliable model that can automatically classify chest CT scan images into two categories: pneumonia and non- pneumonia. For this purpose, Convolutional Neural Networks (CNNs) are utilized due to their proven effectiveness in image classification tasks. The model leverages transfer learning and fine-tuning techniques using pre-trained architectures, enabling better performance even with limited medical datasets. In addition, data augmentation methods such as rotation, zooming, flipping, and shifting are applied to enhance the diversity of the dataset and reduce overfitting. Various optimizers, including RMSprop and Adam, are implemented and compared to improve the training efficiency and accuracy of the model. Experimental results demonstrate that the proposed system achieves high accuracy and strong performance in distinguishing between infected and healthy lung images.

DOI: https://doi.org/10.5281/zenodo.19221433

Nanothermites: Applications, Limitations, Safety Considerations, And Future Prospects

Authors: Anamikaa, Sonia Chaliaa, Manish Naagara

Abstract: The continuous advancement of energetic materials is critical for modern defence, aerospace propulsion, space exploration, and high-performance pyrotechnic applications. Among these, nanothermites—thermite compositions containing at least one nanoscale component—exhibit superior energy density, rapid reaction kinetics, and high combustion temperatures, making them highly attractive for next-generation energetic systems. Their unique properties, including high linear burning rates, tunable reaction heats, and the ability to form hybrid energetic compositions, enable their use in diverse applications such as propellant formulations, micro- thrusters, and reactive materials. However, their widespread adoption faces several challenges. Issues such as the formation of condensed combustion products, which contribute to two-phase flow losses in propulsion systems, and their heightened sensitivity to electrostatic discharge (ESD) pose significant safety risks. Additionally, storage stability and scalability remain major concerns. This study explores the opportunities presented by nanothermites, critically examines their limitations, and discusses key safety considerations. Finally, future research directions are outlined to address existing challenges and enhance the applicability of nanothermites in both civilian and military domains.

Design And Development Of A Nano Satellite For Atmospheric Analysis

Authors: Mr. Kalaimani.N, Ram Pranav Tej Bollina, Rohith Saripallic, Sahith Thorotud

Abstract: The compact Nanosatellite was designed and developed to carry out Atmospheric analysis by creating a step forward to enhance low-cost CubeSat platforms for Environmental monitoring. The modular 6u CubeSat prototype was designed by considering size and weight constraints for my Nanosatellite (weighs <10kg) application. The standard dimensions for 6u CubeSat are 20x10x34.05 in cm. This work focuses on conducting atmospheric analysis by measuring critical parameters such as temperature, pressure, humidity, air quality, and many other factors, which offers many insights into Environmental Research and Atmospheric conditions. The subsystems involved, i.e., The Arduino R4 Wi-Fi, a Microcontroller, and Raspberry Pi 5, a Microprocessor, worked together as on-board computer subsystems to retrieve data from multiple Sensors and scientific instruments, which are integrated into the payload region of CubeSat. The Lora WAN development module operating in the 865-868 MHz frequency range is equipped to engage telemetry and communication subsystem, paired with a custom-designed turnstile antenna deployment system, which allows reliable long-range communication for CubeSat to communicate with the ground station. The CubeSat is structured using lightweight PETG material, with components manufactured through 3D printing, and also equipped with deployed solar panels, which play a vital role in the electronic power supply subsystem to ensure continuous power supply to the CubeSat. Integrating all these subsystems into 6u CubeSat, i.e., a compact nanosatellite that provides a facilitated platform for Atmospheric Research and Remote sensing Applications, will contribute to more innovations in Earth and Space science Applications.

Machine Learning In Money Laundering Detection Over Blockchain Technology

Authors: Mr.B. Mohan, Marthala Muni Namitha, Palloli Veera Vyshnavi, Neelam Venkata Vamsi

Abstract: Layering through cryptocurrency transactions represents a sophisticated mechanism for laundering money within cybercrime circles. This process methodically merges illegal funds into the legitimate financial system. Blockchain technology plays a crucial role in this integration by facilitating the quick and automated dispersal of assets across various digital wallets and exchanges. Machine learning emerges as a powerful tool for analyzing and identifying illicit transactions within Blockchain networks; however, a significant challenge remains in the form of a gap in advanced pattern recognition algorithms. This paper introduces a novel machine learning-based approach called Value-driven-Transactional tracking Analytics for Crypto compliance (VTAC) for the detection of illegal crypto transactions via Blockchain. The approach combines machine learning algorithms with a pre-training process, normalization, model training, and a de-anonymization process to analyze and identify illicit transactions effectively. Experimental evaluations show VTAC’s capability to detect illegal transactions with a 97.5% accuracy using the XG Boost model, outperforming existing methods with an accuracy of up to 95.9%. Key performance metrics, including precision, recall, and F1-score, consistently exceeded 95%, highlighting VTAC’s enhanced precision and reliability. The proposed solution will serve as an advisory framework to help financial crime investigators enhance the detection and reporting of suspicious cryptocurrency transactions in cyberspace.

DOI: https://doi.org/10.5281/zenodo.19230558

InJobs – Interconnected Network Job Search

Authors: Abhiram Krishna A, Adithya A J, Amina H, Daya Deepu, Professor Nitha L Rozario

Abstract: The demand for part-time employment among stu- dents has increased significantly due to rising educational costs and the need for financial independence. However, existing job platforms are not designed to meet student-specific requirements such as flexibility, entry-level accessibility, and security from fraudulent job postings. This paper presents InJobs, an AI- driven web platform that integrates job recommendation, fraud detection, and skill development into a unified system. The platform utilizes BERT-based semantic matching for accurate job recommendations and a Random Forest classifier for fraud detection. Additionally, a learning and certification module is incorporated to bridge the gap between skill acquisition and employment. Experimental evaluation demonstrates improved recommendation relevance, high fraud detection accuracy, and efficient system performance. The proposed system aims to provide a secure, intelligent, and student-focused employment ecosystem.

Smart Planner: A Progressive Web Application For Student Time Management And Academic Productivity

Authors: Akinmerese Oluwatobi, Ifekandu Chiamaka, Obibi Kevin, Iyiade Ayoade

Abstract: This paper presents the Smart Planner System, a Progressive Web App (PWA) to address the inefficiencies in time management and academic productivity faced by university students. Grounded in Covey’s Time Management Matrix for task prioritization and Zimmerman’s Self-Regulated Learning (SRL) Theory for self-regulation, the system integrates a clean user interface to minimize cognitive load. It was developed using Firebase for real-time authentication and Firestore for task management, alongside HTML, CSS, and Vanilla JavaScript, the system features user authentication, a weekly timetable, and customizable notifications. The implementation was detailed through code snippets and a modular architecture. It was evaluated through informal usability testing with peers, confirming intuitive navigation and functionality. Despite limitations such as the absence of formal metrics due to time constraints, the project contributes to educational technology by offering a theoretically informed, student-centered tool. Future enhancements, including Google Calendar integration, gamification, and formal testing, are proposed to further its impact.

DOI: https://doi.org/10.5281/zenodo.19232121

Sentiment Analysis for Social Media Platform

Authors: Srushti Gawade, Rutika Shelke, Manasi Patil, Vaibhav Joshi, Prajkta Jadhav

Abstract: The exponential growth of social media platforms has resulted in massive volumes of user- generated textual data, making manual sentiment interpretation increasingly inefficient and impractical. Sentiment analysis has emerged as a critical natural language processing (NLP) task for extracting meaningful insights from such unstructured content. This research proposes an automated, scalable, and platform-independent sentiment analysis framework designed for social media environments. The current implementation focuses on YouTube comment analysis, where the system collects user comments through the YouTube Data API, performs comprehensive text preprocessing, and applies the TextBlob-based sentiment classification model to categorize comments into positive, negative, and neutral sentiments. In addition to polarity detection, the system incorporates complaint pattern identification and AI-driven suggestion generation to provide actionable insights for content creators and analysts. An interactive visualization dashboard built using Chart.js presents statistical summaries and sentiment distributions to support data-driven decision-making. Experimental evaluation demonstrates that the proposed system efficiently processes large-scale comment datasets while maintaining reliable classification performance suitable for real-world applications. Unlike many existing solutions that are platform-specific, the proposed architecture is modular and extensible, enabling future integration with other social media platforms such as Twitter (X), Instagram, and Facebook. The system has potential applications in digital marketing, brand monitoring, educational feedback analysis, and social media analytics. Future work will focus on incorporating multilingual support, transformer-based deep learning models, real-time streaming analysis, and enhanced emotion detection capabilities. The proposed research contributes toward transforming raw social media feedback into structured business intelligence through an automated and scalable AI-driven approach.

Artificial Intelligence In Cancer Diagnosis, Prediction, And Treatment: A Comprehensive Study

Authors: Ms. Meenakshi, Dr. Brij Mohan Goel

Abstract: This paper will present a comprehensive overview of cancer biology, including its causes, progression, and global statistical trends. It will further examine conventional diagnostic and treatment practices, highlighting their limitations in terms of accuracy, time consumption, and dependency on clinical expertise. The study will then explore how AI technologies, such as Machine Learning (ML), Deep Learning (DL), Natural Language Processing (NLP), and Computer Vision, will enhance the efficiency and reliability of cancer diagnosis and prognosis. In addition, this paper will analyze various cancer data repositories, including radiographic, genomic, pathological, and clinical datasets, which will serve as the foundation for AI-based systems. It will also discuss the emerging applications of AI in oncology, such as early cancer prediction, automated diagnosis, precision medicine, and drug discovery. Furthermore, the research will address key technical challenges, including data scarcity, model interpretability, and generalizability, along with ethical concerns such as data privacy, bias, and accountability in AI-driven decisions. Finally, the paper will emphasize the future potential of AI in transforming cancer healthcare by enabling faster, more accurate, and personalized treatment strategies, thereby improving overall patient outcomes.

DOI: https://doi.org/10.5281/zenodo.19232434

Web Portal For Surgical Products Sales And Services

Authors: Dr. S Malathi, Mr. Madhankumar P

Abstract: The rapid advancement of digital technologies has significantly transformed the healthcare industry, creating a growing demand for efficient, transparent, and integrated systems for managing medical equipment procurement and servicing. Traditional procurement systems often involve fragmented processes, manual documentation, and lack of real-time tracking, which leads to inefficiencies, delays, and increased operational costs. This project presents the design and development of a web-based portal for surgical product sales and service management, aimed at streamlining these processes into a unified digital platform. The system is developed using Python with the Flask framework and SQLAlchemy for database management, ensuring scalability, flexibility, and secure data handling. It integrates core e-commerce functionalities such as product browsing, cart management, order placement, and invoice generation, along with service management features like maintenance request submission, status tracking, and document handling. The platform supports role-based authentication to differentiate between users and administrators, ensuring controlled access and enhanced system security. By combining procurement and servicing operations into a single interface, the system reduces manual workload, improves operational efficiency, and enhances transparency in the healthcare supply chain. This solution demonstrates how modern web technologies can effectively address real-world challenges in medical equipment management.

DOI: https://doi.org/10.5281/zenodo.19248501

House Price Prediction Using Linear Regression And Random Forest Models

Authors: Radha Rani, Ishan Rathi, Preeti Rani, Muskan Thakur

Abstract: The accurate prediction of house prices is crucial for stakeholders in the real estate sector, financial institutions, and urban planners. It not only informs investment decisions but also aids in policy formulation and market analysis. This research paper delves into the comparison of two prominent predictive analytics techniques-Linear Regression and Random Forest—to ascertain their effectiveness in forecasting house prices. Using a Kaggle dataset, this study analyzes key house price predictors such as building classification, living area size, construction year, and land area. The analysis shows Random Forest outperforms Linear Regression in accuracy, emphasizing the importance of building classification and living area in price prediction. Detailed visualizations, like feature importance graphs and scatter plots, offer clear insights into model performance. This research contributes significantly to real estate predictive analytics, offering insights to guide investment strategies and policy-making. It also opens avenues for exploring alternative machine learning approaches and socio-economic factors for a more comprehensive understanding of housing market dynamics.

DOI: https://doi.org/10.5281/zenodo.19248675

An Analysis Of Work–Life Balance Practices And Their Impact On Employee Retention

Authors: K Sridivya, M.Jyothi Prasad

Abstract: The paper aims to provide an in-depth analysis of work-life balance (WLB) practices and their effects on employee retention, based on recent empirical research conducted in different organizational and geographical settings. The research aims to examine the effects of WLB practices on employee retention, job satisfaction, and organizational commitment through an in-depth analysis of WLB practices, such as flexible work arrangements, remote work, supervisor support, and empathy-based benefits. The research aims to develop an integrated work-life balance and retention framework (IWLBRF) that includes direct effects, mediators, and moderators of work-life balance practices on employee retention. The research indicates that effective implementation of work-life balance practices decreases turnover intentions and enhances productivity, particularly in Malaysia, Indonesia, and India. The research indicates that remote work arrangements are not significant in productivity, but they are highly effective in reducing turnover intentions. The research highlights that generation Z is an important factor in work-life balance, as 50% of generation Z consider work-life balance the most critical factor in accepting a job offer, except for salary. Ergonomic practices and empathy-based benefits such as increased bereavement leave and structured flexibility programs demonstrate strong positive relationships with retention, moderated by perceptions of organizational support for personal well-being. The comparative evaluation across the five analytical dimensions—flexible work arrangements, remote work policies, supervisor support, wellness benefits, and generational preferences—indicates that effective retention practices necessitate a holistic approach to WLB.

Demand Response Participation Of Data Centers: Technical Mechanisms, Market Integration, Governance Implications, And Research Directions

Authors: Nimaful Samuel, Hanyabui Augustine, Holison Faith Esther

Abstract: Demand response (DR)—the intentional modification of electricity consumption in response to grid conditions or price signals—has expanded from emergency curtailment into a portfolio of market-based services that can support reliability, integrate variable renewable energy, and reduce infrastructure costs (U.S. Department of Energy, 2006; International Energy Agency, 2023). At the same time, data centers have become a fast-growing and highly concentrated source of electricity demand, with U.S. data center electricity use rising to an estimated 176 TWh in 2023 (4.4% of U.S. electricity consumption) and projected to reach roughly 325–580 TWh by 2028 under scenario assumptions driven in part by accelerated AI server adoption (Shehabi et al., 2024). This co-evolution—rapid load growth and increasing need for flexibility—makes data center participation in DR both attractive and complex. This paper synthesizes peer-reviewed research, national laboratory field studies, standards documents, and selected utility/market operator materials to explain (a) what “data center DR” means in operational terms, (b) how data centers technically deliver flexibility, (c) how DR is communicated, controlled, and verified, (d) how participation is monetized across tariffs and capacity/ancillary markets, (e) what empirical case studies indicate about feasibility and outcomes, and (f) what regulatory, reliability, cybersecurity, and environmental considerations constrain or enable scaling. Key findings supported by the reviewed evidence include: (1) measurable, automation-friendly load flexibility exists in both cooling/infrastructure systems and IT workloads, but the magnitude and response time vary significantly by mechanism and facility type (Ghatikar et al., 2012; Wierman et al., 2014). (5) Recent deployments illustrate two emerging archetypes: “compute-aware DR” (shifting non-urgent tasks across time and sometimes geography) and “grid-interactive UPS/battery services” (fast frequency response and reserve-type services), with high-profile examples reported by large operators and aggregators in multiple regions (Google Cloud, 2023; Microsoft, 2022; Baringa Partners LLP, 2023).

DOI: https://doi.org/10.5281/zenodo.19253383

Design and Implementation of a Rotating Bridge for Train, Ships, And Vehicle Transport at Same Elevation

Authors: Akhilesh Kumar Singh, Km Mansi Mishra, Ankit Kumar, Om Prakash Pal

Abstract: In low lying coastal zones and crowded urban corridors, railway lines, roads, and navigable waterways often compete for space, creating complex infrastructure bottlenecks. Conventional fixed bridges with high clearance demand long approach ramps that conflict with strict railway grade restrictions. Building separate structures for each mode escalates both land use and financial costs. This work proposes a rotating (swing) bridge that accommodates trains, ships, and road vehicles on the same level without any gradient. The structure uses a rim bearing slewing system to rotate a steel truss deck by 90°, allowing ships to pass without vertical obstruction while preserving a flat surface for rail and road traffic. Finite element simulations confirm that the cantilever deflection when the bridge is open stays within L/400. A programmable logic controller (PLC) combined with laser guided positioning yields a rail alignment accuracy of ±1.5 mm, exceeding the required safety margin. The entire opening closing cycle takes less than six minutes, demonstrating practical operability. Compared to a traditional high level bridge, this design reduces steel consumption by about 30% and avoids lengthy approach embankments, making it a cost effective, space saving option for modern multimodal hubs. The results also offer a reference for future smart movable bridges that incorporate artificial intelligence for predictive maintenance and sustainable materials.

An Explainable Hybrid AI System For Multi-Class Brain Tumor Detection Using VGG16 And Large Language Models

Authors: Amarnath, Vivek Upadhyay, Pratyush Dutta Shukla, Mr. Sameer Awasthi

Abstract: Early detection of brain tumors is crucial for improving treatment outcomes, but the diagnosis process is still heavily dependent on specialists’ expertise in analyzing MRI images. This can be limited by high workload, subjective variations in interpretation, and a lack of well-trained professionals [1],[2]. Recently, deep-learning-based automated approaches have achieved high performance in tumor detection, outperforming traditional radiomics approaches in their capability of extracting features directly from images with excellent generalization across various tumor types [3]–[5]. Despite these advances, many of them only output a classification result, failing to provide clinically useful insights into their decisions, which limits their practical usage in clinical settings. In this paper, we present a hybrid diagnostic support system that integrates transfer learning in MRI classification with an explanation module powered by AI. For the classification model, we fine-tuned the VGG16 CNN on a carefully selected dataset of MRI images for glioma, meningioma, pituitary tumor, and no tumor. Taking an inspiration from the recent work conducted on multimodal AI systems, combined with explainable medical imaging systems [6, 7], the proposed system incorporates a large language model, Groq LLaMA-3.3. Our system is trained with explanations in a manner so that it could be clinically adequate to interpret symptoms and provide initial guidance based on the prediction made by the model. This combination helps overcome the lack of transparency often seen in CNN-based medical systems while maintaining a high level of diagnostic accuracy. Results from our experiments reveal that the proposed VGG16 model performs well, matching the effectiveness of other leading CNN models used in brain tumor classification [3], [4], [8]. Adding the module for medical explanation makes the system easier to use, providing predictions in natural language as other AI-driven clinical systems currently under development are doing [7]. In conclusion, this system represents a useful, low-cost, easy-to- understand tool for early screening, conceived to assist-not replace-health professionals, particularly in those geographical areas where radiological competence is lacking.

DOI: https://doi.org/10.5281/zenodo.19254308

Handwriting Recognition System Using OCR

Authors: Mrs Apoora Mane, Miss Anushree Kalloli, Miss Ishwari Kamble, Mr Amol Kote, Mr Tejas Padwal

Abstract: This project presents an Image Text Recognition and Translation System that extracts text from images and converts it into editable and translatable digital content. The system uses image processing techniques to enhance image quality and improve text detection accuracy. By integrating Tesseract OCR, the application efficiently recognizes printed and partially handwritten text from images. After extraction, the recognized text is translated into different languages using an integrated translation module, making the system useful for multilingual communication. Additionally, the system stores the original and translated text in a database, enabling users to maintain a history of their data for future reference. This project aims to reduce manual effort, improve productivity, and provide a user-friendly solution for text extraction and translation. It can be applied in areas such as document digitization, education, and travel assistance. Future improvements may include enhanced handwriting recognition, voice output, and mobile application support.

DOI: https://doi.org/10.5281/zenodo.19254811

Machine Learning For Crop Price Prediction: A Study Of Agricultural Forecasting And Market Analysis Applications

Authors: AJMAL M, ANAS A

Abstract: Agriculture plays a crucial role in the global economy, and accurate crop price prediction is essential for farmers and stakeholders to make informed decisions. Traditional price analysis methods are often manual, unstructured, and lack predictive capabilities, leading to financial risks and inefficient planning. In recent years, machine learning (ML) techniques have gained significant attention due to their ability to analyze large datasets and uncover hidden patterns. This study presents AgriPulse, a machine learning-based web application designed for crop price prediction and market analysis. The system utilizes historical crop price data, rainfall information, and Wholesale Price Index (WPI) values to train a Decision Tree Regression model. It provides six-month and twelve-month forecasts, along with features such as trend visualization, top gaining and losing commodities, and crop profiling. The application is developed using Python and Flask for backend processing, with Pandas, NumPy, and Scikit-learn for data handling and machine learning, while the frontend uses HTML, CSS, JavaScript, and Chart.js for visualization. The proposed system transforms raw agricultural data into actionable insights, helping users optimize decision-making and reduce risks. The study highlights the effectiveness of integrating machine learning with web technologies to enhance agricultural forecasting systems.

DOI: https://doi.org/10.5281/zenodo.19277911

Federated Learning-Based Social Media Analytics

Authors: Shah Yashvi, Harsora Meshva, Shah Archi, Prof. Harkishan Gohil

Abstract: Federated Learning (FL) is transforming social media analytics by enabling privacy- preserving data analysis across distributed platforms. However, traditional analytics methods face major challenges due to data privacy concerns and centralized data collection. FL addresses these issues by allowing model training without sharing raw user data, making analytics more secure and reliable. This paper presents a review of FL in social media analytics, focusing on its importance, techniques, and applications. FL methods such as secure aggregation and differential privacy help analyze user engagement, content trends, and creator performance while protecting user data. These approaches also reduce risks related to privacy, bias, and ethical concerns. Implementing FL in social media analytics helps build user trust, ensures compliance with regulations, and improves data-driven decision-making.

DOI: https://doi.org/10.5281/zenodo.19278294

AgriSmart: AI-Based Smart Farming & Marketplace System

Authors: Prof. Rupnarayan .V.R, Swapnil Bandgar, Abhinav Wadkar, Shivrudra Mangire, Om Dhumal

Abstract: Agriculture is a vital sector for economic development, yet it faces challenges such as low productivity, lack of market access, and dependency on intermediaries. To overcome these issues, AgriSmart: AI-Driven Smart Farming and Digital Marketplace System is proposed as an integrated digital solution that utilizes Artificial Intelligence and web technologies. The system combines key features including a digital marketplace, AI-based crop disease detection, chatbot assistance, and a data analytics dashboard. The marketplace allows farmers to directly connect with customers, eliminating middlemen and ensuring fair pricing. The AI-based crop scanner helps in early detection of plant diseases and provides suitable recommendations for treatment. Additionally, the chatbot (AgriBot) offers real-time guidance related to farming practices, weather conditions, and market trends, while the community platform enables knowledge sharing among farmers. The dashboard provides insights into sales and performance, supporting better decision-making. Overall, AgriSmart improves agricultural efficiency, communication, and productivity. It offers a cost-effective, user-friendly, and scalable solution, bridging the gap between traditional farming and modern digital technologies.

DOI: https://doi.org/10.5281/zenodo.19279132

Design and Analysis of Compact Microstrip Filtenna for Internet of Things Applications

Authors: G.Srikanth, B.Dimpul, CH.Jagadeesh, B.Surya, M.Vinod Kumar

Abstract: This work presents the design of a compact microstrip filtenna for miniaturized Internet of Things (IoT) devices operating in the Long-Term Evolution (LTE) band. The proposed filtenna integrates a 3rd-order hairpin band-pass filter with a coplanar patch antenna on an FR4 substrate (εr = 4.4, thickness = 0.8 mm), enabling both radiation and interference suppression in a single structure. The filter provides a passband from 2.53 GHz to 2.68 GHz, matching the LTE band with low insertion loss and high transmission. The antenna section achieves good impedance matching and stable radiation performance. Simulation results show a resonant frequency around 2.6 GHz with a return loss better than −27 dB and an omnidirectional radiation pattern suitable for IoT applications. The proposed filtenna offers compact size, reduced circuit complexity, and improved out-of-band interference rejection for reliable LTE-based IoT communication.

Terahertz Metasurface Antenna For Wireless Body Area Network

Authors: Dr. K.N.H.Srinivas, Sk.Rajiya, B.Sindhu, G.Ileep, N.L.Madhavi, M.Lukman

Abstract: This paper presents the design and analysis of a meta surface-based antenna system specially created for Wireless Body Area Network (WBAN) applications operating in the terahertz (THz) frequency range. The antenna is designed to meet the growing needs of high-speed, low-power communication in wearable biomedical devices. By integrating a meta surface, the antenna can precisely control electromagnetic waves, improving performance factors such as gain, bandwidth, and efficiency. Using terahertz frequencies allows for better data rates and smaller device sizes. This work introduces an innovative and practical solution for future body-worn technologies.

Multi-Agent Edge–Cloud Systems For Distributed Quality Monitoring

Authors: Dr. Pankaj Malik, Manvi Verma, Manshi Kumari, Mohd. Aamir, Vaibhav Parihar

Abstract: The increasing adoption of Industry 4.0 technologies has led to the need for efficient and scalable solutions for real-time quality monitoring in distributed manufacturing environments. Traditional cloud-centric systems often suffer from high latency, limited scalability, and network dependency, making them unsuitable for time-critical industrial applications. To address these challenges, this paper proposes a Multi-Agent Edge–Cloud System (MAECS) for distributed quality monitoring, integrating edge computing, cloud intelligence, and autonomous multi-agent coordination. In the proposed framework, edge nodes perform real-time defect detection using deep learning models, while cloud servers handle global analytics, model updates, and long-term optimization. A multi-agent architecture enables decentralized decision-making, dynamic task allocation, and efficient resource utilization across the system. The agents collaborate to optimize latency, accuracy, and energy consumption in heterogeneous environments. Experimental evaluation demonstrates that the proposed MAECS significantly outperforms conventional approaches. The system achieves a detection accuracy of 97.3%, compared to 91.2% in cloud-only systems and 93.5% in edge-only systems. Additionally, the proposed approach reduces processing latency to 35 ms, representing a substantial improvement over 250 ms in cloud-based systems and 80 ms in standalone edge solutions. The results confirm that integrating multi-agent coordination with edge–cloud computing enhances both performance and scalability. The proposed system provides a robust and efficient solution for real-time distributed quality monitoring and has strong potential for deployment in smart manufacturing and other industrial IoT applications.

DOI: https://doi.org/10.5281/zenodo.19284544

Lingo Test Confidence Booster With AI Helping Learners Improve Spoken English Through Real Time AI Feedback

Authors: Rabindra Thapa, Bikee Kumar Sah, Rabindra Kumar Mahato, Suraj Babu Patel, Nidhi Patel

Abstract: Public speaking is a vital skill for academic and professional success, yet numerous learners struggle with glos- sophobic bia, limited vocabulary, and weak delivery. Being training approaches, similar to practices or peer evaluation, give limited feedback and aren’t scalable. This paper introduces Lingo-Test,an AI-powered platform designed to enhance spoken English through real-time, multimodal feedback. The system integrates Automatic Speech Recognition(ASR), Natural Lan- guage Processing(NLP), aspect discovery, and facial expression analysis to estimate both verbal and non-verbal performance. Unlike conventional tools that concentrate only on pronunciation or alphabet, Lingo-Test generates a compound confidence score and individualized recommendations covering ignorance, tone variation, vocabulary precarious, and eye contact. The architec- ture includes modular factors for speech processing, sentiment analysis, and feedback visualization, making it scalable for aca- demic and professional operations. An pilot study with learners indicated measurable advancements in ignorance, confidence, and non-verbal delivery. These results punctuate the eventuality of AI- driven multimodal feedback systems to reduce anxiety, strengthen tone mindfulness, and ameliorate communication chops, situatingLingo-Test as a practical result for education and training.

DOI: https://doi.org/10.5281/zenodo.19285046

A Study On Impact of GST in Tirupur Textile Industry

Authors: Dr. R. Indra, Ms. A. Anushiya

Abstract: This work presents the design of a compact microstrip filtenna for miniaturized Internet of Things (IoT) devices operating in the Long-Term Evolution (LTE) band. The proposed filtenna integrates a 3rd-order hairpin band-pass filter with a coplanar patch antenna on an FR4 substrate (εr = 4.4, thickness = 0.8 mm), enabling both radiation and interference suppression in a single structure. The filter provides a passband from 2.53 GHz to 2.68 GHz, matching the LTE band with low insertion loss and high transmission. The antenna section achieves good impedance matching and stable radiation performance. Simulation results show a resonant frequency around 2.6 GHz with a return loss better than −27 dB and an omnidirectional radiation pattern suitable for IoT applications. The proposed filtenna offers compact size, reduced circuit complexity, and improved out-of-band interference rejection for reliable LTE-based IoT communication.

An Intelligent Phishing Website Detection System Using Machine Learning Algorithms

Authors: Krrish Kumar, Kumar Aryan, Akash Kumar, Kumar Divyanshu, Vinay Kumar Pant

Abstract: Phishing attacks are one of the most popular cyber threats, where attackers design a copy of a genuine website to steal confidential information like usernames, passwords, and bank account details. It is quite difficult for common users to identify genuine and phishing websites, resulting in loss of money and data breaches. This project work presents a Machine Learning-based phishing website detection system that examines the URL of the website as well as its HTML structure. It identifies features like URL length, number of links, forms, scripts and external resources. Different algorithms like Random Forest, Support Vector Machine (SVM), Decision Tree, Naive Bayes, and K-Nearest Neighbours (KNN) are used and compared. Among them, Random Forest gave the best accuracy. The system is automated, accurate, and able to identify new phishing websites, thus improving the security of online users.

E-Tongue for Rasa Identification of Ayurvedic Herbs

Authors: Ayush Ashok Shervegar, Rohan Vikas Bombale, Mohammad Kaif Iftekhar Khan, Parth Bahusaheb Bangar

Abstract: Ayurveda classifies medicinal herbs based on the concept of Rasa (taste), which plays a significant role in determining their therapeutic properties. Traditionally, Rasa identification is performed through human sensory perception, which is subjective, non-quantitative, and varies from person to person. To address this limitation, the present project proposes the design and development of a low-cost Electronic Tongue (E-Tongue) system for preliminary Rasa identification of Ayurvedic herbal extracts. The system integrates a pH sensor, Total Dissolved Solids (TDS) sensor, turbidity sensor, and temperature sensor interfaced with an ESP32 microcontroller. The measured physicochemical parameters are processed using rule-based logic to classify the probable Rasa category. A stepper motor mechanism is incorporated to automate sample handling, and the ESP32’s built-in WiFi capability enables real-time monitoring through a web-based dashboard. The prototype was tested on selected herbal powders such as turmeric, neem, amla, and ginger, and distinct variations in sensor readings were observed corresponding to their expected taste characteristics. The developed system demonstrates an economical, portable, and objective approach to bridging traditional Ayurvedic knowledge with modern embedded and IoT technologies, providing a practical solution for academic research and preliminary herbal analysis.

AI Driven Crypto Currency Trading Bot

Authors: N. Indraneel Reddy, N. Manohar Guptha, P. Jyothi Prakash, M. Teja Pavan, Mrs.R. Mano Ranjani

Abstract: Cryptocurrency markets are highly dynamic and operate continuously, requiring traders to analyze price movements and make decisions in real time. Manual trading methods often fail to respond efficiently to rapid market fluctuations, leading to delayed actions and inconsistent results. This paper presents the design and implementation of a real-time cryptocurrency trading bot that performs market analysis, visualizes price movements, and executes simulated trades using technical indicators. The proposed system utilizes live market data obtained from the Binance public API, candlestick chart visualization, portfolio tracking, and an automated trading strategy based on Exponential Moving Average (EMA) crossover logic. The system is implemented as a web application and focuses on transparency, simplicity, and real-time responsiveness. The results demonstrate that automated rule-based trading can improve decision consistency and provide a practical platform for understanding algorithmic trading concepts

Advanced Security Monitoring System Using LoRa/LoRaWAN By Raspberry Pi

Authors: S. Susmitha, V. Hema Sri Durga, K. Sandeep, M. Bulliraju, T. Sunil Abhishek, R.L.R. Lokesh Babu

Abstract: A comprehensive approach is developed for the design and implementation of a solar-powered, Raspberry Pi-based edge processing perimeter alert system capable of detecting unmanned aerial vehicles (UAVs) and enhancing security monitoring. This system is intended to protect key areas such as borders, private property, and critical buildings. The setup uses a Raspberry Pi equipped with Artificial Intelligence (AI) to recognize drones and helicopter sounds. Including PIR motion sensors, microwave radar, and microphone are integrated to detect movement and changes in the monitored area. For long-distance communication, the system employs LoRa/LoRaWAN technology, which operates with very low power, supports long-ranges, and can transmit signals over few meters. The entire unit runs on a solar-charged battery, enabling operation in remote locations without access to electricity. This makes system eco-friendly and suitable for continuous 24/7 security monitoring. Field testing under controlled conditions demonstrates high reliability and precision in UAV detection, validating the feasibility of combining edge AI processing, renewable energy, and low-power communication technologies for decentralized security applications. This architecture provides a scalable, energy-efficient, and environmentally sustainable solution for aerial threat detection and prevention of unauthorized access.

Edge-Enabled Dual-Mode IoT System For Water And Environmental Monitoring With Smart Alerts

Authors: Gopala Reddy P, Kusuma Naga Sai Sahiti P, Lehya Sri Venkata Harshitha N, Kusuma Satya Sai Jyothi P, Ganesh Pavan Kumar CH, Mohan Sai

Abstract: The problem of water quality deterioration and the environmental pollution is a burning issue that endangers the health of the population, the level of agricultural activity, and the sustainable development. Traditional methods of water quality monitoring are based on manual sampling or cloud-based Internet of Things (IoT) that constantly transmits raw sensor data, thereby causing a greater latency, the increase in network traffic, and a high reliance on the stable internet connection. To overcome these shortcomings, this paper introduces an edge-based dual-mode IoT system of real-time water quality and environment monitoring with smart alert messages.The suggested system incorporates the sensors to measure water quality parameters such as pH, turbidity, total dissolved solids, electrical conductivity and temperature as well as environmental parameters such as air temperature, humidity and atmospheric pressure. The Raspberry Pi Pico 2W is a data processing board that features edge-level data processing with filtering, sliding-window averaging, and threshold analysis to achieve rapid detection of anomalies without constant cloud communication. The two-mode Wi-Fi and GSM communications are used to guarantee the security of data transmission in the areas where the network covers are insufficient. Processed data are visualized through the Blynk IoT platform, while local alerts are generated using RGB LEDs and a buzzer. The system is cost-effective, scalable, and suitable for deployment in both rural and urban environments.

Smart Transformers Health Monitoring Using AI-ML And IoT Integration

Authors: Kalyani, Srinivas, Revanth, Monica, Navya

Abstract: Power Transformers are considered critical and costly equipment in electrical power systems. Sudden transformer failures can cause considerable economic losses. Conventional transformer monitoring systems have used periodic manual inspection for transformer monitoring. Such methods may be ineffective for detecting faults at the primary stage. This paper proposes an IoT-based transformer health monitoring system in real-time using ESP32 and Raspberry Pi Zero 2W. The system monitors various critical transformer parameters in real time, including temperature, gas concentrations (Hydrogen, Methane), oil level, load currents, operating voltage, and ultrasonic fault signals. The ESP32 module will be used to collect the transformer parameters and transmit the values to Firebase for cloud storage. In addition, Machine learning (ML) models like Random Forest, KNN, Gradient Boosting, and Support Vector Machine (SVM) are used for fault prediction. The collected data is sent to Firebase, where a web-based dashboard built with ReactJS provides visualization and overall analysis of the system. This system enables early fault detection and extends the transformer’s lifespan.

A Smart Approach On The Collecting Working Condition Data From Home Appliances Under The Field Test

Authors: Teliki Sainath, Vaggu Ganesh, Venugonda Shameerbasha, Teliki Sainath, Vaggu Ganesh

Abstract: This research presents an intelligent method for managing and collecting operational data from home appliances during the field test process. The proposed approach involves deploying sensors to gather device-specific and environmental data, which are then systematically evaluated to assess the working conditions of each appliance. The data management system ensures efficient storage, interpretation, and transmission of information, enabling real-time monitoring and analysis. The methodology not only facilitates predictive maintenance and energy management but also supports the identification of possible faults and optimization of appliance performance in actual user environments. The smart approach contributes to the advancement of smart home systems by integrating IoT-based sensor networks and data analytics to improve reliability, energy efficiency, and user experience.

DOI:

A Study Of Competativeness Of Organized Retail Sector In Tamilnadu With Reference To Online/E Retail Stores

Authors: Ms. Sri Vidya R, Ms.R.Sowmya

Abstract: The retail sector in India has undergone significant transformation over the last decade, with organized retail and e-retail channels emerging as major players in consumer markets. Tamil Nadu, being one of the most industrialized states, exhibits a dynamic retail environment. This study aims to examine the competitiveness of the organized retail sector in Tamil Nadu with special reference to online/e-retail stores. By analyzing consumer preferences, pricing strategies, service quality, and technological adoption, the study highlights key factors driving competition and offers insights for retail managers and policymakers The organized retail sector has witnessed significant transformation over the past decade, largely driven by the growth of online and e-retail stores. With technological advancements and increasing internet penetration, consumer shopping behavior has shifted from traditional brick-and-mortar stores to digital platforms. This shift has intensified competition among organized retailers to attract and retain customers through better services, pricing strategies, and innovative offerings. This shift has created a highly competitive environment where retailers must adopt innovative strategies to attract and retain customers.

DOI:

The Coastal Public Domain In Lebanon: Legal Conflicts, Spatial Encroachments, And Strategies For Sustainable Coastal Management

Authors: Dr. Mohamad H. Jichi

Abstract: The Lebanese coastline represents one of the country’s most valuable territorial and environmental assets, providing important opportunities for economic development, tourism, public recreation, and environmental protection. Legally defined as part of the maritime public domain, coastal areas are intended to remain public property accessible to all citizens. However, despite the existence of legal frameworks regulating coastal land use, Lebanon’s coastline has experienced significant transformation over the past decades due to rapid urbanization, tourism development, and weak regulatory enforcement. This study examines the current condition of the coastal public domain in Lebanon and analyzes the major spatial, legal, environmental, and governance challenges affecting its management. The research adopts a qualitative analytical approach based on the review of legal documents, policy reports, and international coastal management practices. Particular attention is given to the impacts of illegal coastal encroachments, privatization of shoreline areas, loss of public access to the sea, and environmental degradation of coastal ecosystems. The findings reveal that the Lebanese coastal zone suffers from fragmented governance structures, insufficient enforcement of maritime public domain regulations, and the absence of an integrated coastal planning strategy. As a result, large segments of the coastline have become inaccessible to the public and environmentally degraded. The study highlights the importance of adopting Integrated Coastal Zone Management (ICZM) principles to improve coastal governance in Lebanon. It proposes a set of planning and policy recommendations including the establishment of a national coastal management strategy, strengthening legal enforcement mechanisms, implementing coastal zoning regulations, restoring public access to the sea, and promoting sustainable coastal tourism development. By reorganizing the coastal public domain through integrated planning and governance reforms, Lebanon can transform its coastline into a sustainable national resource that supports economic growth, environmental protection, and social equity.

DOI: https://doi.org/10.5281/zenodo.19344390

Facial Recognition Attendance Monitoring System

Authors: Sanyam Mittal, Vanshika Garg, Aaditya Jain, Shubhi Verma

Abstract: Traditional attendance systems employed in educational institutions and workplaces suffer from inherent inefficiencies, including susceptibility to proxy attendance, high administrative overhead, and slow data processing. This paper presents the design and implementation of an automated Facial Recognition Attendance Monitoring System (FRAMS) developed using Java and the OpenCV computer vision library. The proposed system leverages the Haar Cascade Classifier for robust real-time face detection and the Local Binary Pattern Histogram (LBPH) algorithm for accurate face recognition. The architecture integrates a webcam-based image acquisition module, a preprocessing pipeline for noise reduction and face normalization, an LBPH-trained recognition engine, and a MySQL database for persistent attendance storage. Experimental results demonstrate a recognition accuracy of up to 97.4% under optimal lighting conditions, with an average frame processing time of 210 milliseconds. The system effectively eliminates proxy attendance, reduces administrative workload, and enables real-time monitoring without requiring specialized hardware. Evaluation across diverse environmental conditions confirms the system’s robustness, with performance metrics substantially outperforming conventional attendance modalities. This work contributes a practical, cost-effective, and scalable solution to institutional attendance management.

Smart Mirror

Authors: Shruti Deshmukh, Prachi Sonkawade, Rutuja Waghmare, Sara Nikam

Abstract: A Smart Mirror is an innovative device that combines a traditional mirror with modern digital technology to provide useful information to users. The main purpose of a smart mirror is to display real-time data such as time, date, weather updates, news headlines, and calendar reminders while functioning as a normal mirror. This project uses components such as a two-way mirror, display screen, Raspberry Pi, and internet connectivity to create an interactive system. The smart mirror works by placing a display behind a two-way mirror, where the Raspberry Pi processes and shows information collected from the internet. Users can view important updates while performing daily activities like grooming or getting ready, which saves time and improves convenience. Some advanced smart mirrors may also include features like voice commands, sensors, and smart home control. The Smart Mirror project demonstrates the practical use of Internet of Things (IoT) technology in everyday life. It aims to make daily routines more efficient by integrating digital information into a common household object.

VISISCAN – AI-Driven Smart Visitor Management

Authors: P.Sunitha Devi, Daya Varshini, Nikhita Makam, M.Kavya Sree, R.Ashwitha Reddy

Abstract: Visitor management is an important tool in the context of providing security and transparency in running of institutions. The conventional methods such as manual registers tend to be ineffective and subject to mistakes. VisiScan is an AI- based Smart Visitor Management System that offers the solution of using facial recognition and AI-based automation to secure and contactless check-in. The efficiency is also increased by incorporating real-time tracking, voice-enabled feedback, and encrypted cloud storage into the system as well as role-based access and automated notifications provided to administrators. VisiScan fosters institutional openness and is in line with SDG 9 and SDG 16 which enhances innovation and strong institutions. Controlling visitor access in an efficient and secure manner is one of the necessary requirements in present-day institutions and organizations. The conventional visitor management software, e.g. manual registers and physical passes, frequently results in the inaccuracy of the data, time delays and security issue. All these approaches lack real-time updates, adequate verification, and maintenance of records on a long-term basis and hence are unsuitable to the modern safety and operational requirements. The Digital transformation has enabled the visitor management process to be automated and enhanced with the rapid development of the Artificial Intelligence (AI) and facial recognition. The use of AI-based systems can guarantee high speed and accuracy of authentication, less human activity, and less administrative overhead. Cloud-based databases also enhance the availability of data and security of storage, which allows organizations to have the right and open records of visitors.

AI-Driven Project Collaboration and Team Management Platform

Authors: Mukesh Kumar Kasireddy, N Nikhil Kumar Reddy, N Srinivas, Dr. Sasikumar Gurumoorthy

Abstract: Modern student projects require collaboration among members with different technical skills such as frontend development, backend development, database management, artificial intelligence, and UI/UX design. However, students often face challenges in forming effective teams, coordinating tasks, managing communication, and integrating project components using multiple disconnected tools. This paper proposes an AI-Driven Project Collaboration and Team Management Platform designed to provide a unified environment for student teams to collaborate efficiently. The platform integrates team formation, real-time communication, task management, document sharing, and project monitoring into a single system. Artificial intelligence is incorporated to recommend compatible teammates based on skills and interests, extract tasks from discussions, summarize conversations, and provide code suggestions. The proposed system utilizes a modern web architecture consisting of a React.js frontend, Spring Boot backend, PostgreSQL database, and WebSocket-based real-time communication. AI services powered by large language models are used to analyze chat data and assist in project management. The platform improves productivity, reduces coordination overhead, and provides students with an industry-like collaborative development environment.

Design And Analysis Of Glass Fiber Reinforced Polymer (GFRP) Leaf Spring

Authors: Ashwin M. Patil

Abstract: Weight reduction is now the main issue in automobile industries. Weight reduction can be achieved primarily by the introduction of better material, design optimization and better manufacturing processes. The achievement of weight reduction with adequate improvement of mechanical properties has made composite a very good replacement material for conventional steel. Selection of material is based on cost and strength of material. The composite materials have more elastic strain energy storage capacity and high strength to weight ratio as compared with those of steel, so multi-leaf steel springs are being replaced by mono-leaf composite springs. The paper gives the brief look on the suitability of composite leaf spring on vehicles and their advantages. The objective of the present work is design, analysis and fabrication of mono composite leaf spring. The design constraints are stress and deflections. The material selected is glass fibre reinforced plastic (GFRP) and the epoxy resin can be used which is more economical to reduce total cost of composite leaf spring with similar mechanical and geometrical properties to the multileaf spring. The composite leaf spring is fabricated by hand lay-up technique and tested. The testing was performed experimentally with the help of UTM and by (FEA) using ANSYS software showing stresses and deflections were verified with analytical and experimental results. Compared to the steel spring, the composite spring has stresses that are much lower, the spring weight is nearly 74% lower.

Autism Prediction Using Deep Learning

Authors: Tharuni R, Tamilarasi S, Sathiya V, Subedha V, Sharmila S, Jabasheela L

Abstract: Healthcare is essential for human survival. The term “autism disease” encompasses a broad spectrum of symptoms utilized for diagnosis. Techniques for assessing diseases early on helped figure out the best way to handle high- risk people, which lowered their risk. The main goal is to keep people safe by spotting strange behavior. Researchers are working on ways to predict autism. Which disease can be diagnosed early? The model requires enhancement. In this paper, we propose a unified model which is hybrid of CNN and Bi-LSTM models to use deep learning methods to detect the presence of autism in individuals. We tackle the issues of missing and unbalanced data in the by employing data processing methods.

DOI: https://doi.org/10.5281/zenodo.19364751

Data-Driven Customer Segmentation & E- Commerce Retail Optimization

Authors: Vivek Vinod Prasad, Surapuraju Jagadeeswar Raju, Rohit Lakhamanbhai Vala

Abstract: The growth of e-commerce businesses is heavily reliant on data-driven decision-making, where customer segmentation and personalized marketing strategies play pivotal roles. This paper proposes a framework for leveraging machine learning and data analytics to optimize e-commerce operations, focusing on customer segmentation and churn prediction. We explore the application of the Random Forest algorithm for predicting customer churn, and RFM (Recency, Frequency, Monetary) analysis for effective customer segmentation. The proposed system also incorporates Customer Lifetime Value (CLV) calculations to forecast the potential revenue from each customer, aiding businesses in resource allocation. By combining machine learning models with a robust Flask-based platform, the system enables real-time analytics, personalized product recommendations, and automated insights for administrators and customers. Security features, including role- based access control, ensure secure data management. Through this framework, e-commerce businesses can enhance customer engagement, reduce churn, and drive revenue growth by making informed, data-backed decisions.

DOI: https://doi.org/10.5281/zenodo.19368353

“JARVIS: AI- Based Voice Assistant System”

Authors: Prof.Suryawanshi .A.P, Ashutosh Hakale, Gujar Priyanshu, Chavan Vedant

Abstract: In today’s digital era, voice-based technologies are transforming the way users interact with computer systems. However, traditional interaction methods such as keyboards and mice are still widely used, which require manual effort and time. To overcome these limitations, this project presents JARVIS: Voice Assistant System, an intelligent platform that uses speech recognition and artificial intelligence to perform tasks through voice commands. The proposed system integrates multiple functionalities such as voice input processing, command execution, real-time responses, and web-based interaction. The system captures user voice, converts it into text, processes the command using Python, and provides output either in text or voice format. JARVIS can perform various operations such as opening applications, searching the internet, providing time and date information, and answering general queries. The system is developed using HTML, CSS, and JavaScript for frontend and Python for backend processing. The results show that the system improves user interaction, reduces manual effort, and provides a faster and more efficient way of performing tasks. It is a cost-effective, user-friendly, and scalable solution for modern intelligent systems.

DOI:

The Major Obstacles Encountered by Secondary School Student’s in Comprehending Specific Topics in Chemistry.

Authors: Vester Hanyinde, Dr. Sumathi. K. Sripathi

Abstract: This study investigates the obstacles secondary school students face in comprehending specific topics in chemistry a pressing issue that affects academic performance and future career prospects. The methodology employed involves a comprehensive review of the current curriculum, teaching methods and assessment techniques, alongside the implementation of innovative and interactive approaches to learning, such as hands-on experiments, simulations and multimedia resources. The study utilized both written and oral interviews. It was found that schools lacking learning materials, laboratories, and technology perform poorly in chemistry. Data were analyzed using bar charts and pie charts. The implementation of this methodology led to improved student engagement, motivation and academic achievement, as well as reinforcing better preparation for the future careers in STEM (Science, Technology, Engineering and mathematics). In conclusion, the scope of this is far reaching. Addressing it is crucial to ensuring that secondary school students develop a deep understanding of chemistry and its applications which is essential for driving innovation, economic growth and sustainable development.

DOI: https://doi.org/10.5281/zenodo.19369280

Legal Documents Summarizer

Authors: Hariharasudhan N, Harish k, Harish Ragavendar N, Mrs.P.G.Gayathri

Abstract: The manual analysis of legal instruments—ranging from binding contracts and complex agreements to judicial precedents—is often impeded by their intricate syntax and voluminous nature. This research presents an intelligent, automated summarization framework designed to distill lengthy legal texts into concise, actionable summaries without compromising semantic integrity. The proposed system employs a multi-layered Natural Language Processing (NLP) pipeline, incorporating rigorous preprocessing phases such as lemmatization and tokenization alongside domain-specific cleaning. Distinguishing itself from traditional tools, this architecture utilizes a hybrid methodology that integrates graph-based ranking (TextRank/LexRank) for extractive precision with Transformer-based models (BART/Legal-BERT) for abstractive coherence. Furthermore, the system incorporates a conversational interpretation module powered by Retrieval-Augmented Generation (RAG) to allow interactive clause clarification. Initial findings suggest that this dual-model approach significantly enhances document accessibility and professional efficiency, bridging the gap between complex legal terminology and user comprehension.

DOI: https://doi.org/10.5281/zenodo.19383787

Digital Twin Based Microgrid Monitoring And Control System

Authors: R.Ganeshram, V.Kamal Govindhan, R.Sri Hari Prasath, Mr.P.Tamilnesan

Abstract: The increasing need for energy systems that provide dependable and sustainable energy has caused a surge in the growth of microgrids utilizing renewable energy sources. However, there are still many difficulties related to the implementation of distributed energy resources, such as real-time monitoring and effective control of energy resources. In response to those difficulties, the developed system is a Digital Twin Based Microgrid Monitoring and Control System that uses an ESP32 controller with Wi-Fi capability to connect all components of the microgrid to a cloud-based interface (as the digital twin) and a MATLAB environment for analysis and visualization of real-time data received from the monitored electrical parameters such as voltages, currents and power. The Digital Twin model represents (virtually) the physical microgrid system and therefore allows the user to evaluate the system performance, this can include detecting faults and developing optimized methods for controlling any of the distributed energy resources. As a result of using the Digital Twin Based Microgrid Monitoring and Control System, the microgrid will be more reliable, the operating losses will be less and it will support the implementation of smart grids for future sustainable energy management solutions.

DOI: https://doi.org/10.5281/zenodo.19384084

Determinants Of AI Acceptance Among Health Insurance Consumers In Jammu & Kashmir: Examining The Mediating Role Of Trust And The Influence Of Socio-Economic Factors

Authors: Rohan Sharma, Dr. Rohit Bhagat

Abstract: This paper aims to provide a conceptual model to understand the factors affecting the acceptance of artificial intelligence (AI) technology among health insurance consumers in Jammu & Kashmir. Though health insurance organizations are embracing AI technology at a rapid pace in claims processing, consumer interactions, and fraud detection, consumer adoption of AI technology is low compared to organizational adoption. This paper aims to understand the relationship between socio-economic factors such as income, education, digital literacy, geographic location, and consumer trust as a mediating factor affecting AI technology adoption through a study of empirical data collected between 2021 and 2026. A Contextual Trust Mediated Acceptance Model (CTMAM) has been developed to understand consumer adoption of AI technology in a specific socio-cultural context such as J&K, where infrastructural challenges have shaped consumer technology adoption behaviors in a unique manner compared to other regions. Analysis of empirical data indicates that though 94% of health insurers are actively adopting AI technology, only 21% of health insurance consumers are adopting AI technology, where trust acts as a mediating factor. Concerns for privacy (20%), accuracy perceptions (26%), and transparency issues have a significant impact on the acceptance of AI. Socio-economic factors play an important role in the relative importance of these issues. The comparative analysis of the demographic segments shows that the acceptance of AI is lower for the rural population and women due to the low level of trust in digital technologies.

DOI: https://doi.org/10.5281/zenodo.19385765

Wildlife And Poaching Detection System

Authors: Ayush Saini, Akshat Saini, Ajay, Ms. Dipti Dhiman

Abstract: Wildlife conservation has become a critical global concern due to increasing incidents of illegal poaching and habitat destruction. Traditional monitoring methods are inefficient and resource-intensive, making it difficult to ensure continuous surveillance of vast wildlife reserves. This research proposes an AI-based Wildlife and Poaching Detection System using the YOLO (You Only Look Once) object detection model for real-time monitoring. The system integrates computer vision, deep learning, and web-based deployment to detect animals and unauthorized human activities from video feeds. The model is trained on diverse datasets and achieves high detection accuracy under varying environmental conditions. The system provides instant alerts, reduces dependency on manual patrolling, and enhances conservation efforts. The proposed solution is scalable, efficient, and suitable for deployment in real-world wildlife protection scenarios.

A Study On Complementary Products To Change In Economic System

Authors: Gowtham.A. G, Dr.M.D Chinnu

Abstract: This study examines the role of complementary products in influencing changes in the econom-ic system. Complementary products are goods that are jointly demanded and consumed to-gether by consumers. The research aims to analyse consumer awareness and purchasing be-haviour related to such products. Primary data for the study was collected through a struc-tured questionnaire using Google Forms. The study focuses on how price changes in one product affect the demand for its complementary product. The findings indicate that comple-mentary products significantly influence demand and market growth. The technology sector shows the strongest interdependence among complementary goods. The study highlights the importance of complementary products in shaping modern economic systems. The research concludes that complementary products play a crucial role in economic development and poli-cy decisions.

“WanderLust: A Web-Based Travel Planning Platform”

Authors: Nareandra Sanjay Bhute, Rohan Bhanudas Pawar, Rutik Nandkishor Pawar, Aditya Madan Mapari, Prof. N. N. Kumbhar

Abstract: The rapid proliferation of online travel services has resulted in a fragmented ecosystem where travelers must navigate multiple disconnected platforms to book accommodation, rent vehicles, and discover dining options. This paper presents the design and implementation of WanderLust, a full-stack, multi-category travel booking platform that consolidates stays, vehicle rentals, and dhaba (local dining) reservations into a single unified web application. The system is built on a Node.js v18 and Express.js v4 backend with MongoDB Atlas as the NoSQL database, employing a Model-View-Controller (MVC) architecture with EJS server-side rendering. Key implemented features include: (i) an AI-powered trip planning module and conversational chatbot leveraging the OpenAI GPT-4o-mini API with function calling for real-time database queries; (ii) dual payment gateway integration with Stripe (international) and Razorpay (Indian UPI, net banking) alongside an in-app digital wallet; (iii) real-time host–guest communication via Socket.IO WebSockets; (iv) geospatial proximity search using Mapbox SDK with MongoDB 2dsphere indexing; (v) a comprehensive security layer comprising Helmet.js Content Security Policy, CSRF token validation, rate limiting, XSS sanitization, and NoSQL injection prevention; (vi) a multi-channel notification system spanning email (Nodemailer/Brevo API), browser push notifications (Web Push), and in-app alerts; and (vii) cross-platform mobile deployment via Capacitor for native Android packaging. The platform manages 14 Mongoose data models, 24 route modules, and 20 utility services. Authentication is handled through Passport.js with local credentials and Google OAuth 2.0, augmented by OTP-based email verification. Automated testing was conducted using Jest and Supertest with an in-memory MongoDB instance. The application is deployed on Render (PaaS) with Cloudinary CDN for media delivery and Sentry for real-time error monitoring. Evaluation demonstrates a feature-complete, production-grade platform that addresses the identified research gap of no existing unified open-source solution integrating multi-category bookings, AI-assisted itinerary generation, dual-gateway Indian payment support, and hybrid mobile deployment within a single cohesive architecture.

DOI:

Leaf Disease Detection Web Application Using Deep Learning

Authors: Aslam Amir Shaikh, Borate Sukeshkumar Pandurang, Mule Tanuja Suresh, Deshpande Yogesh Narendra

Abstract: Agriculture is under serious threat, and this threat includes diseases that affect plant leaves. Our method pinpoints both the disease that affected the leaf and the region of damage. Both the quantity and quality of agricultural products are impacted by crop diseases, particularly those that primarily harm the leaves. The human eye’s ability to see subtle differences in the diseased leaf area is not as strong as it should be. In this study, we provide an automated web-based system for classifying and diagnosing plant leaf diseases. We are using CNN model as a feature extractor and prediction model to swiftly categorise illnesses. Proper treatment can be provided by the study of disease. Along with this we have utilized the docker, POSTMAN API and TF serving server to make the system scalable and improve the working. This study is validated using the Plant Village dataset for plants like tomato and potato. The training and testing results indicate that the CNN model have a greater classification accuracy than the currently in use ANN model. The proposed approach could prove a useful tool for farmers and industry specialists to utilise when making decisions about crop management and disease control.

DOI: https://doi.org/10.5281/zenodo.19387083

IdeaLoop: A Modern Full-Stack Discussion Forum

Authors: Nisha Kumari, Mr. Sachin Mishra

Abstract: IdeaLoop is a modern, full-stack web-based discussion forum designed to promote effective online community interaction. The platform is developed using Next.js 14 with React Server Components, ensuring high performance, scalability, and seamless user experience. IdeaLoop enables users to authenticate securely through GitHub OAuth 2.0, create discussion topics, publish posts, and participate in structured, nested comment threads. The system follows a Reddit-style discussion model, allowing hierarchical conversations with unlimited comment depth for better engagement and clarity. Prisma ORM with SQLite is used for efficient database management, while NextAuth.js handles secure authentication and session management. Tailwind CSS and Radix UI are used to create a responsive and accessible user interface compatible with both mobile and desktop devices. Additionally, the platform integrates real-time search functionality to enhance content discovery. By leveraging server-side rendering and modern web technologies, IdeaLoop delivers a reliable, secure, and scalable solution for online discussions, making it suitable for educational, professional, and community-based applications.

Deep Learning For Stock Market Prediction: A Long Short-Term Memory (LSTM) Approach To NSE Tata Global Beverages Limited Closing Price Forecasting

Authors: Samarth Tyagi, Surya Verma, Pratyaksh Garg, Dr. Nitin Gupta

Abstract: Stock market prediction remains one of the most challenging and consequential problems in computational finance. This paper presents a comprehensive deep learning framework leveraging Long Short-Term Memory (LSTM) recurrent neural networks for time-series forecasting of equity closing prices. Using five years of historical trading data (2013–2018) from NSE-listed Tata Global Beverages Limited, comprising 1,235 trading observations, we construct a stacked dual-layer LSTM architecture trained on a 60-day lookback window with MinMax normalization to prevent data leakage. Our model—trained over 5 epochs with a batch size of 2 and the Adam optimizer—achieves convergence with a Mean Squared Error (MSE) loss of approximately 9.06 × 10⁻⁴. On an 80/20 train-validation split (987/248 observations), the model demonstrates strong temporal alignment between predicted and actual closing prices. The experimental results validate LSTM’s effectiveness in capturing long-range sequential dependencies in financial time-series data, outperforming traditional statistical models in non-stationary market environments. This work contributes a reproducible, modular pipeline for equity price forecasting with practical implications for algorithmic trading, portfolio management, and financial risk modeling.

DOI:

Face Detection System: A Comprehensive Study

Authors: Ankit Kr. Karn, Nabin Sah, Battu Akash, Ajay Kumar Sah

Abstract: Face detection has emerged as a cornerstone task in modern computer vision, forming the backbone of numerous real-world applications, including biometric authentication, security and surveillance systems, smartphone unlocking features, and social media tagging. Over the past two decades, significant advancements have been made in this field, evolving from early handcrafted feature-based algorithms to sophisticated deep learning architectures capable of handling complex scenarios. Despite these advancements, designing a robust and efficient face detection system that performs reliably under diverse conditions remains a challenging problem. This project aims to design and implement a comprehensive face detection system by integrating both traditional and modern methodologies. The proposed approach involves careful dataset selection, rigorous preprocessing, and the application of classical techniques such as Haar Cascade alongside modern convolutional neural network (CNN)-based frameworks, including Multi-task Cascaded Convolutional Networks (MTCNN) and YOLO-based detectors. Through this methodology, the project evaluates the performance, accuracy, and real-time efficiency of different detection strategies. The results of the study demonstrate that the system achieves high efficiency in real-time detection while effectively identifying faces under varying poses, scales, and illumination conditions. However, certain challenges remain, including handling occlusions, extreme pose variations, and low-light scenarios, which continue to affect detection accuracy. The project concludes by suggesting future directions for improvement, such as incorporating bias mitigation strategies, exploring multimodal biometric systems, and implementing liveness detection to enhance security.

Bone Fracture Detection Using Deep Learning-Based Medical Image Analysis

Authors: Nitin Kumar, Ishita Singhal, Dr. Nitin Gupta

Abstract: Bone fractures represent one of the most prevalent medical conditions resulting from traumatic events such as accidents, falls, and sports-related injuries. Accurate and timely detection of fractures is critical for effective treatment and patient recovery. Conventional diagnostic methods rely heavily on manual interpretation of X-ray images by radiologists, which can be time-consuming and susceptible to human error, particularly in cases involving subtle fracture patterns or large volumes of medical data. This paper presents the design and implementation of an automated Bone Fracture Detection System utilizing deep learning and computer vision techniques. The proposed system employs Convolutional Neural Networks (CNNs) for feature extraction and classification, along with advanced architectures such as VGGNet for image classification and Faster R-CNN for fracture localization. The system is trained on a dataset comprising both fractured and non-fractured X-ray images, enabling it to learn complex visual patterns associated with bone abnormalities. The implementation is carried out using Python and integrates powerful libraries including TensorFlow, Keras, OpenCV, and NumPy. A user-friendly web-based interface is developed using Streamlit, allowing users to upload X-ray images and obtain real-time predictions. The system processes input images through preprocessing techniques such as normalization, resizing, and augmentation before performing classification and detection tasks. Experimental results demonstrate that the proposed system achieves high accuracy in detecting bone fractures, with reliable performance across varied image conditions. The integration of classification and object detection models enables both identification and localization of fractures, enhancing the interpretability of results. The system significantly reduces diagnostic time and supports healthcare professionals in decision-making processes. This work highlights the potential of deep learning in medical image analysis and provides a scalable, efficient, and cost-effective solution for automated fracture detection. The proposed system can serve as a valuable decision-support tool, particularly in resource-constrained healthcare environments.

DOI: https://doi.org/10.5281/zenodo.19396111

Multi-Tank Batch Process System

Authors: Dr. Vipul Shah, Pathik Patel, Tanish Hande

Abstract: In small chemical industries we instrumentation people have to focus on control the temperature of on going processes and also focus on other field parameter, Temperature & Level plays very crucial role in batch chemical manufacturing industry.This research describes how we can control the level and also the main parameter Temperature, So for that we have do some research on how to make small scale project where we can learn both field devices and also about controllers.Using PLCs, we can control this batch process for live controlling and monitoring of our process, Also for real-time monitoring we use SCADA system. For industrial safety, device safety, & also for Human safety we are controlling this all field parameters and for that field device plays a crucial role, here an Instrumentation & Control Engineer monitors the process on SCADA system. And if needed than we can change the parameters values and makes it easier and safer for real-time controlling industry. Instrumentation & control engineers are not only focuses on field devices they are also responsible for safety of field devices and also controllers safety, This safety is given by an automation engineer they are also an IC people , For perfect automation we have mainly know about our application & according to that we are selecting controllers, supplies & also safety devices like MCB, RELAY, CONTACTOR, VFDs, Terminal Blocks(TB) etc.

AUDS: Design And Development Of Agentless Unified Defense System

Authors: Dr. Jagdish W. Bakal, Paras Thakur, Darin Joy Peringalloor, Vaishnavi Pawar

Abstract: Today most of the internet users faces different types of cyber threats like phishing websites, DNS spoofing attacks, malicious redirects, and also unwanted tracking activities. However users make the use of blacklists and signature based detection in order to block the websites but this method is not effective for newly created malicious websites. So to solve this issue, we decided to design and developed the Agentless Unified Defense System (AUDS). This system gives real time protection without any need of extra software. It works directly inside our browser. When the site looks unsafe, then it blocks the access. Also when the site appears to be suspicious, it opens in a separate Disposal Window to keep the browser safe. In this way, AUDS provides simple web security.

DOI: https://doi.org/10.5281/zenodo.19397160

Multi-Tank Batch Process System

Authors: Dr. Vipul Shah, Pathik Patel, Tanish Hande

Abstract: In small chemical industries we instrumentation people have to focus on control the temperature of on going processes and also focus on other field parameter, Temperature & Level plays very crucial role in batch chemical manufacturing industry.This research describes how we can control the level and also the main parameter Temperature, So for that we have do some research on how to make small scale project where we can learn both field devices and also about controllers.Using PLCs, we can control this batch process for live controlling and monitoring of our process, Also for real-time monitoring we use SCADA system. For industrial safety, device safety, & also for Human safety we are controlling this all field parameters and for that field device plays a crucial role, here an Instrumentation & Control Engineer monitors the process on SCADA system. And if needed than we can change the parameters values and makes it easier and safer for real-time controlling industry. Instrumentation & control engineers are not only focuses on field devices they are also responsible for safety of field devices and also controllers safety, This safety is given by an automation engineer they are also an IC people , For perfect automation we have mainly know about our application & according to that we are selecting controllers, supplies & also safety devices like MCB, RELAY, CONTACTOR, VFDs, Terminal Blocks(TB) etc.

Aerodynamic Drag Reduction Of A Passenger Car Model Using CFD Analysis In ANSYS Fluent

Authors: Aarthi S, Amirdha Varshni S, Janani Kavya R, Jayabharathi R, Dr. S. Santhanakrishnan

Abstract: In this world as there is increment growth in automobile industry also there is need to develop a high efficient vehicle which can take high speed performance with high stability. An improvement in high speed of a vehicle depends upon amount of drag force subjected to a vehicle and it has adverse effect of increment in energy consumption of the vehicle. Since the testing using Wind tunnel is expensive and time consuming, it is required to develop some effective numerical methods. The paper states to provide a CFD based aerodynamic analysis of an object which could be useful for determining and reduction of drag force for the passenger car by the application of ANSYS Fluent software. Critical aerodynamic quantities to be observed for such a system are the distributions of pressure, the contours of velocity and the wake and streamlines in relation to a vehicle. To provide an efficient prediction such critical quantities are given with relevant turbulence models and boundary conditions as velocity inlet, pressure outlet and no slip wall condition. Pre-processing for such a problem can be defined as geometry creation, domain creation and meshing. The obtained results based on the experiment show that the redesign of the geometry has reduced the value of coefficient of drag. Such CFD methodology to provide cost effective and efficient car analysis prior to any physical prototype can be implemented in latest car designing.

DOI: https://doi.org/10.5281/zenodo.19399656

Comprehensive Survey Of Advanced Time Series Signal Processing

Authors: Tushar Parulekar, Sandeep Chilukuri

Abstract: Time series data is observed in daily activities ranging from financial markets and automotive sensors to the medical industry and weather prediction. Proper analysis of this data plays a pivotal role in the era of artificial intelligence; with correct interpretation, we can utilize the data to its full potential. This article provides a holistic survey of the state of the art in time series signal processing, spanning from classical spectral decomposition and statistical filtering to the application of foundation models. We evaluate various architectures, including Convolutional Neural Networks (CNNs), Transformers, Graph Neural Networks (GNNs), and the emerging class of Structured State Space Models (SSMs) such as Mamba, specifically regard- ing their application to time series data. Additionally, we provide an overview of signal processing within deep learning contexts, exploring hybrid frameworks comprising wavelet transforms, Fourier analysis, and Kalman filtering. Finally, we assess the challenges faced in applying these concepts to time series data and discuss obstacles regarding the deployment of lightweight models.

DOI: https://doi.org/10.5281/zenodo.19400712

Synthesis, Spectral And Biological Studies Of Transition Metal Complexes Of 4-(2-(3-fluorobenzylidene)hydrazinyl)-7H-pyrrolo[2,3-d]pyrimidine

Authors: Mr. Hanumant Rananaware, Dr. M. A. Badgujar

Abstract: A series of transition metal complexes of 4-(2-(3-fluorobenzylidene) hydrazinyl)-7H-pyrrolo[2,3-d]pyrimidine (L) were synthesized by reacting the ligand with metal salts of Fe(III), Cu(II), Co(II), Ni(II), and Zn(II) under reflux conditions. The complexes were characterized by various spectroscopic techniques, including UV-Vis, FTIR, NMR, and X-ray diffraction, which confirmed the coordination of metal ions to the ligand through the azomethine nitrogen and pyrimidine nitrogen. The electronic spectra revealed significant shifts in the absorption maxima upon complexation, indicating coordination-induced electronic changes. FTIR and NMR analyses further supported the bidentate nature of the ligand. Magnetic susceptibility and conductivity measurements suggested that the complexes exhibit low to moderate conductivity and a range of magnetic properties, consistent with the coordination environment of the metal ions. The antibacterial activity of the ligand and its metal complexes was evaluated against Gram-positive (Staphylococcus aureus) and Gram-negative (Escherichia coli) bacteria using the disk diffusion method. The metal complexes exhibited superior antibacterial activity compared to the free ligand, with the Cu(II) and Co(II) complexes demonstrating the highest inhibition zones, particularly against E. coli. The increased activity of the metal complexes is attributed to enhanced membrane penetration and metal-ligand synergy, which likely alters the microbial cell's normal physiological processes. The results suggest that these complexes hold promise as potential antimicrobial agents, and further studies are warranted to explore their full therapeutic potential.

DOI: https://doi.org/10.5281/zenodo.19414795

Synthesis, Structural Characterization, And Biological Evaluation Of Salicyaldehyde-Derived Fluorinated Chalcones With Enhanced Antimicrobial Activity

Authors: Poonam Kumari, C S Azad, Niranjan Kumar Mandal

Abstract: A series of novel fluorinated chalcone analogues incorporating a hydroxybenzaldehyde framework was successfully synthesized and systematically investigated. The synthetic approach involved a base-catalyzed Claisen–Schmidt condensation between various hydroxy-substituted benzaldehydes and fluorinated acetophenone derivatives. The structures of the synthesized compounds were unequivocally established using detailed spectroscopic techniques, including ^1H NMR, ^13C NMR, infrared spectroscopy, and mass spectrometry. The antimicrobial activity of the prepared chalcone derivatives was evaluated against a panel of clinically relevant bacterial and fungal strains, including Staphylococcus aureus, Escherichia coli, Pseudomonas aeruginosa, and Candida albicans. Several compounds demonstrated significant antimicrobial potency, with minimum inhibitory concentration (MIC) values ranging from 5 to 25 μg/mL. Structure–activity relationship (SAR) studies revealed that the antimicrobial efficacy is strongly influenced by the number and positional arrangement of hydroxyl groups, as well as the electronic effects imparted by fluorine substitution within the chalcone scaffold. In conclusion, the study highlights the potential of fluorinated hydroxychalcone derivatives as promising antimicrobial candidates and provides important insights for the rational design and development of more potent next-generation antimicrobial agents.

DOI: https://doi.org/10.5281/zenodo.19416569

Addressing Grid Interconnection And Power Infrastructure Challenges For Sustainable Data Center Expansion In The United States

Authors: Samuel N Nimaful, Augustine Hanyabui, Faith Esther Holison, Laureta Tatenda Nyamsutswa, Joel Holison

Abstract: U.S. data center growth—driven by cloud services, artificial intelligence (AI), and digitalization—has become a power-system planning problem as much as a real-estate or information-technology problem. Recent federal analysis indicates data centers consumed about 176 terawatt-hours (TWh) in 2023 (about 4.4% of total U.S. electricity use) and could reach roughly 325–580 TWh by 2028 (about 6.7%–12%), depending on broader economic and electricity demand growth (U.S. Department of Energy [DOE], 2024). [1] These levels of load growth are material at a bulk-power-system scale and are arriving faster than most traditional generation, transmission, and distribution (T&D) investment cycles—especially in regions already constrained by transformer lead times, siting and permitting timelines, and limited interconnection and construction capacity. [2] Interconnection bottlenecks—historically framed as a generator problem—now collide with large-load connection timelines and cost responsibility debates. At the transmission level, DOE reports the interconnection queues have expanded from under 500 gigawatts (GW) in 2010 to roughly 2,600 GW “today,” with more than 95% associated with zero-carbon generation and storage; DOE also notes that time-to-interconnect has more than doubled nationally. [3] Independently, Berkeley Lab’s queue synthesis reports that, as of end-of-2024, thousands of projects and thousands of gigawatts remain in queues, and median durations from interconnection request to commercial operation have risen to over four years for recently completed projects in regions with available data. [4] Even when data centers are not themselves “in the queue” (because loads often follow different or less transparent processes than generation), the same constrained equipment, studies, and network upgrade construction resources govern whether new load can be served on feasible timescales. [5] Power infrastructure constraints are now observable in reliability performance. NERC identifies “large loads”—including data centers—among the most significant near-term reliability challenges, citing observed events in which approximately 1,500 MW of data centers disconnected simultaneously and unexpectedly from the bulk electric system (BES) in 2024 after a transmission fault, creating balancing and stability challenges analogous in magnitude to a large nuclear plant changing output unexpectedly. [6] This operational reality re-frames “speed-to-power” strategies: getting connected quickly is not sufficient if poor load observability, protection coordination, and flexibility arrangements raise system risk or shift costs to other customers. [7]

DOI: https://doi.org/10.5281/zenodo.19422330

River Depth Monitoring Using GPS Integrated Cleaning Robot

Authors: Mr. Ravishankar B V, Mr. Siddarth S, Mr. Vishal, Mr. Yashwanth H V, Mr. Yashwanth. B, Dr. M J Anand

Abstract: This project presents the design and development of an River Depth Monitoring Using GPS Integrated Cleaning Robot, an innovative river monitoring and maintenance solution. The system integrates a GPS- guided navigation platform, sonar-based depth measurement, and hazard detection for real-time mapping and environmental analysis. Key features include autonomous waypoint navigation, data communication via Bluetooth, onboard data logging, and a mobile application interface for user interaction. Sustainability is emphasized with waterproofing, optional solar power, and eco-friendly materials. Enhanced safety through obstacle detection and emergency recovery ensures reliable performance, making this a versatile tool for environmental monitoring and river maintenance.

IMGCRYPT: Secrets Hidden Within The Ordinary

Authors: Professor Gurav P.S., Nirbhay Patil, Saksham bansode, Sahil Bhujbal

Abstract: In today’s digital world, securing sensitive information is very important. Traditional encryption protects data but may attract attention. To overcome this, steganography hides the existence of data itself. This paper presents IMGCRYPT, a system that combines image steganography and encryption to securely hide confidential information inside ordinary images. The proposed system ensures that the hidden data remains invisible and protected even if detected. The approach uses encryption algorithms along with Least Significant Bit (LSB) technique to embed data into images. The system is efficient, secure, and user-friendly, making it suitable for secure communication.

An AWS-Driven Intelligent Framework For Scalable Data Deduplication And Storage Optimization

Authors: Professor ,Dr. Y. Jayababu1, Chikkala Kedareswari Kaivalya, Darna Mahathi3, Siddanthapu Sai Sri Ram4, Kalepu Dhanush Sai5

Abstract: Cloud storage systems are experiencing rapid growth due to the increasing demand for scalable and cost-effective data management solutions. However, redundant data storage leads to excessive storage consumption, increased bandwidth usage, and higher operational costs. This paper proposes a data deduplication framework using Amazon Web Services (AWS) to eliminate duplicate files in cloud environments. The system generates a unique hash value using the MD5 hashing algorithm whenever a file is uploaded to Amazon S3. AWS Lambda functions are used to compute and compare hash values stored in Amazon DynamoDB to identify duplicate files. If a duplicate file is detected, the system prevents redundant storage and maintains a reference to the original file, thereby optimizing storage utilization. Experimental evaluation demonstrates improved storage efficiency, reduced memory consumption, and stable server response time. The proposed approach enhances cloud storage performance while maintaining data integrity and security.

DOI: http://doi.org/10.61463/ijset.vol.14.issue2.175

Deep Learning-Based Intelligent Fire Detection And Early Warning System Using Computer Vision

Authors: Assistant Professor,Mrs.T.Satya Aruna, Vendra Sri Harsha, Dadala Anusha, Manchimsetti Lokesh, Gubbala Gnana Satya Sai, Rajala Jaya Surya

Abstract: Early detection of fire is critical for preventing large-scale disasters, minimizing property damage, and ensuring public safety. Traditional fire detection systems mainly rely on physical sensors such as smoke, heat, and gas detectors. Although these methods are widely used, they often suffer from delayed detection, high false alarm rates, and limitations in complex environments such as industrial facilities and crowded urban areas. With the rapid advancement of computer vision and deep learning technologies, intelligent image-based fire detection systems have emerged as an effective alternative for improving early fire detection. This paper proposes a deep learning-based intelligent fire detection and early warning system that uses Convolutional Neural Networks (CNN) to automatically identify fire in images captured from surveillance cameras. The proposed system analyses visual features from images and classifies them into two categories: fire and non-fire. A structured dataset containing fire and non-fire images is used to train and validate the deep learning model. Data augmentation techniques such as image rotation, scaling, and horizontal flipping are applied to improve model generalization and reduce overfitting. In addition, training optimization techniques including Early Stopping and ReduceLROnPlateau are implemented to enhance model performance and stability. Experimental results demonstrate that the CNN-based model significantly outperforms traditional machine learning techniques such as Logistic Regression, K-Nearest Neighbor (KNN), and AdaBoost. The proposed model achieves high classification accuracy while maintaining strong recall and AUC performance metrics. Furthermore, the system integrates an automated alarm mechanism that generates an alert when fire is detected, enabling rapid emergency response. The proposed approach provides a cost-effective, reliable, and scalable fire detection solution that can be deployed in surveillance systems for buildings, industrial environments, and smart city infrastructures. The results indicate that deep learning-based visual fire detection systems can significantly enhance disaster prevention and safety monitoring capabilities

DOI: http://doi.org/10.61463/ijset.vol.14.issue2.176

Machine Learning-Enhanced Post-Quantum Cryptographic Framework For Secure Data Protection In The Quantum Computing Era

Authors: Assistant Professor ,Mrs.S.Surya Sri, Pasagodugula Ramya, Sigala Gowthami Ganeswari, Katari Srinivas Durga Mahesh, Bangaru Mahendraraj, Kondeti Sugnan Raj

Abstract: The rapid advancement of quantum computing poses a significant threat to traditional cryptographic systems that rely on computational hardness assumptions such as integer factorization and discrete logarithms. Quantum algorithms, particularly Shor’s algorithm, have the potential to break widely used cryptographic standards including RSA and Elliptic Curve Cryptography. To address these emerging challenges, this research proposes a machine learning-enhanced post-quantum cryptographic framework designed to provide adaptive and resilient data security in the quantum computing era. The proposed framework integrates post-quantum cryptographic algorithms with intelligent machine learning mechanisms to improve the robustness of encryption systems against quantum-based attacks. The system incorporates three major components: a post-quantum encryption module, a machine learning–based anomaly detection module, and an adaptive key management system powered by reinforcement learning. The anomaly detection component utilizes neural network models to monitor encrypted data streams and identify abnormal patterns that may indicate potential decryption attempts. In addition, reinforcement learning dynamically adjusts cryptographic key parameters and rotation schedules to reduce the risk of key compromise. Experimental evaluation demonstrates that the proposed framework significantly improves detection accuracy, reduces key management latency, and enhances resilience against simulated quantum decryption attacks when compared with conventional cryptographic systems. The integration of machine learning enables the framework to dynamically adapt to evolving threats while maintaining strong encryption standards. The results highlight the potential of combining artificial intelligence with post-quantum cryptography to build adaptive, intelligent, and future-proof security systems capable of protecting sensitive data in the presence of powerful quantum computing technologies.

DOI: http://doi.org/10.61463/ijset.vol.14.issue2.177

A Hybrid Neural Network And Xgboost Framework For Accurate Disaster Prediction And Management

Authors: Assistant Professor Mrs.T.N.V.Durga, Appanapalli Santosh, Chintakayala Hima Bindu, Mohammad Anjum Sharifa, Pilla Sai Manikanta, Pakalapati Sanjay Varma

Abstract: Natural disasters such as floods, wildfires, and earthquakes cause severe damage to human life, infrastructure, and the global economy. As the frequency and intensity of these disasters continue to increase due to climate change and environmental factors, the need for reliable disaster prediction systems has become more critical than ever. Traditional disaster prediction methods often rely on historical patterns and statistical models, which are limited in their ability to handle complex and imbalanced datasets.This study proposes a hybrid machine learning framework that integrates Neural Networks and XGBoost to improve the accuracy and reliability of disaster prediction. In the proposed approach, neural networks are used to automatically extract meaningful and high-level features from disaster datasets, while XGBoost performs the final classification of disaster types. To address the common issue of class imbalance in disaster datasets, the Synthetic Minority Over-sampling Technique (SMOTE) is applied during data preprocessing.The model is evaluated using a real-world disaster dataset containing records of floods, wildfires, and earthquakes. Experimental results demonstrate that the proposed hybrid model significantly outperforms traditional machine learning techniques such as Random Forest, Support Vector Machines, and Logistic Regression. The proposed system achieves high prediction accuracy and improved F1-scores across all disaster categories.Overall, the hybrid Neural-XGBoost framework provides a robust and efficient solution for disaster prediction and management. The system can support disaster management agencies by enabling early warning systems, improving preparedness strategies, and assisting in better resource allocation during emergency situations.

DOI: http://doi.org/10.61463/ijset.vol.14.issue2.178

Hybrid Transfer Learning And Machine Learning Framework For Accurate Food Image Classification

Authors: Assistant Professor ,Mr.S K Sankar, Yeluri Vidhya Dhari, Kommana Pavan Vinaykumar3, Madala Komala Sri Anjani Patnaik, Malladi Kasubabu, Nethula Santhosh Kumar

Abstract: The rapid growth of digital food datasets and the increasing demand for automated dietary monitoring systems have created a need for efficient food classification techniques. Traditional machine learning approaches often struggle to handle the high-dimensional and complex nature of food images, mainly due to limitations in manual feature extraction and scalability. To address these challenges, this research proposes a hybrid framework that combines transfer learning models with machine learning classifiers for accurate food image classification.In the proposed approach, pre-trained deep learning architectures are utilized to extract meaningful visual features from food images, enabling the system to capture complex patterns and representations. Models such as EfficientNet, DenseNet, and MobileNet are employed as feature extractors due to their strong performance in image recognition tasks. The extracted feature vectors are then classified using machine learning algorithms including XGBoost and Random Forest to improve prediction accuracy and interpretability.The hybrid framework integrates the feature learning capability of deep neural networks with the decision-making efficiency of classical machine learning algorithms. Experimental evaluation demonstrates that this combination improves classification accuracy and robustness, even when dealing with noisy or diverse food image datasets. The results indicate that the proposed system can effectively classify multiple food categories and can be applied in real-world applications such as nutritional monitoring, automated dietary assessment, and food safety management systems.

Machine Learning Driven Audio Signal Analysis For Automated Hate Speech Detection In Short-Form Social Media Videos

Authors: Assistant Professor, Mr.Y.Manas Kumar,, Iragavarapu Sri Vishnu Chittha Priya, Pepakayala Bhuvaneswari, Nookala Sri Giridhara Nageswara Adity, Koduri D P V Sai Sruthi, Yalla Naya Samson

Abstract: The rapid growth of social media platforms has significantly increased the spread of harmful and offensive content, including hate speech. Short-form video platforms allow users to share content quickly, making it challenging to monitor and control abusive speech. While most existing hate speech detection systems rely heavily on textual analysis, many harmful expressions occur through spoken language in video content. This study presents a machine learning-based approach for detecting hate speech directly from audio signals extracted from short-form online videos. The proposed framework collects audio data from publicly available social media videos and processes the signals using several audio feature extraction techniques such as Mel Frequency Cepstral Coefficients (MFCC), Spectral Centroid, Spectral Rolloff, Spectral Bandwidth, Zero Crossing Rate, and Chroma features. These features are used to train supervised machine learning models including Logistic Regression, Support Vector Machine, and Random Forest classifiers. To ensure reliable evaluation, a 5-fold cross-validation strategy is employed along with performance metrics such as accuracy, precision, recall, and F1-score. Experimental results demonstrate that the Random Forest model achieves superior performance compared to other classifiers by effectively capturing important audio characteristics associated with hate speech patterns. The study highlights the significance of spectral features and MFCC representations in identifying hateful expressions in speech. The proposed approach provides a practical framework for automated monitoring of harmful audio content in modern social media platforms and can contribute to improving online content moderation systems.

DOI: http://doi.org/10.61463/ijset.vol.14.issue2.180

Intelligent Climate Trend Prediction Using Hybrid Machine Learning Models And Spatiotemporal Data Analytics

Authors: Associate Professor ,Dr.D.Uma, Koppili Gnana Keerthana, Kona Lakshmi Laalasa, Madarapu Bulli Raju, Andamani Haswanth Nag, Dunna Guna Shekhar

Abstract: – Climate change has become one of the most critical global challenges, influencing environmental stability, agriculture, and human livelihoods. Accurate analysis and prediction of climate trends are essential for developing effective mitigation and adaptation strategies. Traditional statistical approaches often struggle to capture complex relationships present in large-scale climate datasets. To address this limitation, this study proposes an intelligent climate prediction framework based on machine learning techniques and spatiotemporal data analytics. Historical climate data containing temporal attributes such as year and month along with spatial parameters including latitude and longitude are used to analyse long-term temperature variations. The proposed system applies multiple machine learning algorithms, including Linear Regression, Random Forest, Support Vector Regression, and K-Nearest Neighbor, to identify patterns and predict future climate trends. Data preprocessing and feature extraction techniques are employed to improve model performance and reduce noise in the dataset. Experimental evaluation demonstrates that ensemble-based models provide higher predictive accuracy compared to traditional regression approaches. The results highlight the effectiveness of machine learning models in interpreting climate data and forecasting temperature variations. This research contributes to the development of intelligent climate monitoring systems that can support environmental research, policy planning, and sustainable development initiatives.

Intelligent Disaster Monitoring Using Social Media Data With Location Intelligence And Sentiment Analysis

Authors: Associate Professor, Dr.Sreepada Sarada1,, Pasupuleti Sri Durga Tanuja Gayatri2,, Gurugubelli Jhansi3,, Kandelli Neelanjana4,, , Pappala Akash5, Bonam Sumanth Kumar6

Abstract: In recent years, social media platforms have become an important source of real-time information during natural disasters and emergency situations. Millions of users share posts, images, and location information that can provide valuable insights for disaster monitoring and response. However, identifying relevant disaster-related information from the massive volume of social media data remains a significant challenge. This paper presents an AI-driven disaster detection framework that utilizes social media analytics, location intelligence, and sentiment analysis to monitor and identify disaster events in real time. The proposed system collects social media posts and processes them using natural language processing and machine learning techniques to detect disaster-related content. Location intelligence methods are applied to extract geographical information from posts, enabling accurate identification of affected areas. In addition, sentiment analysis is used to evaluate public emotions and urgency levels associated with disaster events. The integrated framework helps emergency response teams gain situational awareness and make timely decisions during critical situations. Experimental evaluation demonstrates that the proposed approach effectively identifies disaster-related posts and provides meaningful insights for disaster management systems. The framework can support authorities and emergency organizations in improving response strategies and enhancing public safety.

DOI: http://doi.org/10.61463/ijset.vol.14.issue2.182

Deep Contextual Situational Analytics For Intelligent User Behaviour Prediction In Smart Environments

Authors: Professor,Dr.M.Radhika Mani1 ,, Yarlagadda Sai Sukeerthi2,, Kondi Durga Pravallika3,, Guttula Samuel4,, Zainul Fathima5,, Shaiki Ali Hussain Baba6

Abstract: Understanding user behaviour in dynamic environments has become an important challenge in modern intelligent systems. Traditional user modelling approaches often rely on static information and fail to capture the influence of changing contextual situations such as location, time, activity, and environmental conditions. To address this limitation, this paper presents an intelligent context-driven situational analytics framework designed to analyse and model user behaviour in smart environments. The proposed approach integrates contextual data collection, situation identification, and behavioural analysis to dynamically interpret user activities and preferences. By utilizing contextual parameters and adaptive analytical techniques, the framework enables systems to better understand user needs and provide more accurate and personalized responses. The model improves decision making by continuously analysing situational patterns and predicting possible user actions. Experimental evaluation demonstrates that incorporating situational context significantly enhances the effectiveness of user behaviour prediction compared to traditional static user models. The proposed framework can be applied in various smart applications including personalized recommendation systems, intelligent assistants, and context-aware services.

DOI: http://doi.org/10.61463/ijset.vol.14.issue2.183

Crack Vision-AI: A Deep Transfer Learning Framework For Structural Crack Detection Using MobileNet

Authors: Assistant Professor, Mrs.P.Lakshmi Satya1, Shaik Fuzaila Farhatunnisa 2,, Mummidi Naga Durga Venkat3, Mohammad Samreen 4,, Sunkara Rajeev Nagendra5

Abstract: Structural cracks in buildings, bridges, and other civil infrastructures pose serious risks to public safety and long-term durability. Early and accurate detection of cracks is essential for effective maintenance and prevention of structural failures. Manual inspection methods are often time-consuming, labour-intensive, and prone to human error, making automated crack detection systems increasingly important.This project presents a deep learning–based framework for image-based crack prediction using a lightweight Convolutional Neural Network architecture. The proposed system utilizes MobileNet as the backbone model and incorporates transfer learning to enhance feature extraction capabilities while reducing computational complexity. Extensive image preprocessing and data augmentation techniques are applied to improve model generalization and handle limited dataset availability.The model is trained and evaluated on a structured crack image dataset consisting of cracked and non-cracked samples. Performance evaluation is carried out using standard metrics such as accuracy, precision, recall, F1-score, and confusion matrix analysis. Experimental results demonstrate that the proposed approach achieves high classification performance while maintaining computational efficiency, making it suitable for real-time infrastructure monitoring applications.The developed framework contributes toward automated structural health monitoring by providing a reliable, scalable, and efficient crack detection solution adaptable to practical engineering environments

An Explainable AI-Driven Multimodal Deep Learning Framework For Intelligent Android Malware Detection

Authors: Assistant Professor Mr.G.Vijay Kumar1,, Yalla Aishwaryambica2, Pemmada Venkata Vamsi3,, Dasari Deshma Susmitha4,, Mohammad Vazeeruddin5,

Abstract: With the rapid growth of Android applications, malware attacks targeting mobile devices have increased significantly, posing serious security and privacy threats to users. Traditional malware detection techniques, including signature-based and rule-based methods, often struggle to identify newly emerging or obfuscated malware variants. To address these limitations, this study proposes an explainable artificial intelligence-based framework, referred to as XAI-Droid, for effective Android malware detection and classification.The proposed system integrates deep learning techniques with explainable AI (XAI) mechanisms to not only improve detection accuracy but also provide transparent and interpretable decision-making. Feature extraction is performed using static analysis techniques, and the processed features are used to train advanced machine learning and deep learning models. To enhance trust and reliability, explanation methods such as feature importance analysis are incorporated to highlight the key attributes influencing classification decisions.Experimental results demonstrate that the proposed framework achieves high detection accuracy while maintaining interpretability, making it suitable for real-world cybersecurity applications. By combining robust classification performance with explainability, XAI-Droid contributes to the development of trustworthy AI-based mobile security systems.

DOI: http://doi.org/10.61463/ijset.vol.14.issue2.160

Spinexnet: An Intelligent Deep Learning Architecture For Multiclass Spinal Disorder Classification From Radiographic Images

Authors: Associate Professor ,Dr.D.Uma1, Yadavalli Sai Naga Sri Ramya2,, Latchipatula Lohitha3,, Bongu Sai Tulasi Pravallika4, Sabhavathula Pujyavenkata Krishnachaitanya

Abstract: Spinal disorders such as scoliosis, osteoporosis, spondylolisthesis, osteochondrosis, and vertebral compression fractures are major contributors to chronic back pain and physical disability worldwide. Early and precise diagnosis using spine X-ray imaging plays a crucial role in effective treatment planning and long-term patient care. However, manual analysis of radiographic images is time-consuming and highly dependent on the expertise of radiologists, which may lead to variability in diagnosis.To address this challenge, this study presents a deep learning-based multi-class classification framework for the automatic detection of various spinal conditions from X-ray images. The proposed system utilizes Convolutional Neural Networks (CNNs) for automated feature extraction and classification. Multiple pre-trained architectures, including VGGNet and ResNet, are evaluated and compared with a customized CNN model to identify the most effective approach. The dataset undergoes preprocessing steps such as resizing, normalization, and augmentation to improve model generalization and robustness.Experimental results demonstrate that the proposed CNN model achieves superior performance, with high accuracy, precision, recall, and F1-score across multiple spine condition categories. The system provides reliable and consistent predictions, highlighting its potential as a computer-aided diagnostic tool. By assisting medical professionals with faster and more standardized analysis, the proposed framework can contribute to improved clinical decision-making and better patient outcomes

DOI: http://doi.org/10.61463/ijset.vol.14.issue2.159

A Resource-Efficient Deep Neural Framework For Bearing Health Monitoring Using Current Sensor Analytics

Authors: Assistant Professor, Mrs.D.Kanaka Mahalakshmi Devi1, Gunipe Surya Teja2,, Killi Vijaya Vardhan3,, Gadiyakari Karthik4, Jyothula Hareesh5, Bollavarapu Moses Jude Christopher6

Abstract: Fault diagnosis of rolling bearings is a critical task in industrial motor systems, as bearing defects can lead to severe mechanical failures and costly downtime. Traditional vibration-based monitoring systems require additional hardware and complex signal processing techniques, making them expensive and difficult to deploy in practical environments. In contrast, current sensor-based fault diagnosis offers a more economical and convenient alternative, as motor current signals can be collected without installing extra sensors.This project presents LiteFDNet, a lightweight deep learning framework designed for efficient and accurate bearing fault diagnosis using motor current signals. Instead of directly processing high-dimensional raw signals, the proposed approach extracts meaningful time-domain statistical features to reduce computational complexity. A compact neural network architecture with residual and dense connections is implemented to enhance feature representation while maintaining low model complexity. Additionally, explainable feature selection techniques are applied to identify the most informative features contributing to fault classification.Experimental results demonstrate that LiteFDNet achieves high diagnostic accuracy while significantly reducing computational cost and inference time compared to conventional deep learning models. The proposed system is suitable for real-time industrial applications, particularly in resource-constrained edge computing environments

DOI: http://doi.org/10.61463/ijset.vol.14.issue2.158

Marine Vision -AI: A Comparative Analysis Of Machine Learning And Deep Learning Methods For Underwater Marine Species Classification

Authors: Assistant Professor, Mrs.N.Durga Deepti Priya1 , Vulli Venkata Gangadhar Praveen2, Kudipudi Nikhil Vijay Kumar3, Chappidi Siva, Vulli Venkata Gangadhar Praveen2, Kudipudi Nikhil Vijay Kumar3, Chappidi Siva Sai Venkata Hari Suresh4, Galla Pavan Kumar5, Sri Krushna Vamsi Ram6

Abstract: The classification of underwater marine species plays an important role in marine biodiversity monitoring, ecological research, and conservation planning. However, identifying marine species from underwater images is a challenging task due to poor lighting conditions, water turbidity, background noise, and variations in species appearance. Traditional manual identification methods are time-consuming and require expert knowledge, making automated classification systems highly valuable. This project presents a comparative analysis of Machine Learning (ML) and Deep Learning (DL) techniques for the classification of underwater marine species. A real-time underwater image dataset containing 189 images across 20 different marine species is used for experimentation. Traditional machine learning models such as Support Vector Machine (SVM), K-Nearest Neighbors (KNN), and Decision Tree are evaluated alongside deep learning architectures including AlexNet, DarkNet-19, and SqueezeNet. Experimental results demonstrate that deep learning models significantly outperform traditional machine learning methods in terms of classification accuracy. Among all evaluated models, SqueezeNet achieves the highest accuracy, demonstrating its effectiveness in handling complex underwater visual patterns while maintaining computational efficiency. The study highlights the advantages of convolutional neural networks in extracting meaningful features from underwater images and emphasizes their suitability for marine species classification tasks.

DOI: http://doi.org/10.61463/ijset.vol.14.issue2.157

A CNN–Autoencoder-Based Deep Learning Framework For Automated Detection Of Rotator Cuff Tendon Tears In Shoulder Ultrasound Images

Authors: Assistant Professor,Mr.Sayantan Kar1 ,, Karri Pravallika 2,, Damera Rajya Lakshmi Nikshitha 3,, Chintada Venkatesh 4,, Mannam Alekhya Narayana 5 ,, Banala Dineshbabu6

Abstract: Rotator cuff tendon tears are one of the most common causes of shoulder pain and mobility limitations worldwide. Although ultrasound imaging is widely used for diagnosis due to its affordability and real-time capability, accurate interpretation heavily depends on the experience of radiologists. Variability in image quality, speckle noise, and unclear anatomical boundaries often make manual assessment challenging and time-consuming.This project presents an automated deep learning framework for detecting shoulder rotator cuff tendon tears from ultrasound images. The proposed system combines a Convolutional Neural Network (CNN) with an autoencoder-based contour segmentation approach to accurately identify key anatomical structures such as the humeral cortex and subacromial bursa. By focusing on meaningful structural boundaries instead of traditional pixel-wise segmentation, the model achieves improved precision and robustness. The segmented outputs are further utilized for classification using a deep CNN architecture to distinguish between intact and torn tendons.Experimental evaluation demonstrates strong segmentation accuracy and reliable classification performance, highlighting the potential of the proposed method as a supportive diagnostic tool. This system can assist clinicians in making faster, more consistent decisions while reducing dependency on manual interpretation. The approach represents a significant step toward explainable and efficient AI-driven medical image analysis in musculoskeletal ultrasound diagnostics.

DOI: https://doi.org/10.61463/ijset.vol.14.issue2.156

Plant Leaf Diseases Prediction: Using Machine Learning

Authors: Mr. Yogesh N. Deshpande, Dr. Yuvraj V. Parkale, Mr. Aslam A. Shaikh

Abstract: This study investigates the monitoring of progressive compressive strength development in concrete using non-destructive testing (NDT) techniques, aiming to provide a reliable, in-situ alternative to traditional destructive testing. An experimental program was conducted on concrete specimens (mixes with varying strength ranges) to track strength evolution at early and mature ages, specifically 7, 28, and 90 days. The NDT techniques, primarily focusing on Ultrasonic Pulse Velocity (UPV) and the Rebound Hammer (RH) test, were utilized to estimate the compressive strength (SonReb method). The research confirms that while sole NDT methods possess inherent inaccuracies (roughly $pm-20% for RH), combining RH and UPV significantly improves the precision of in-situ concrete strength predictions.This study presents a systematic investigation of the monitoring of progressive compressive strength of M30 Concrete using a pendulum-based non-destructive testing (NDT) technique. The research is based on the principle of the coefficient of restitution, where the rebound angle of a pendulum after impact is correlated with the compressive strength of concrete. Traditional strength evaluation using the Compression Testing Machine (CTM) is destructive and unsuitable for in-situ monitoring. To overcome this limitation, a pendulum-based device is used to measure rebound angle at different curing ages (2, 7, 14, 21, and 28 days).Two empirical relationships were developed 1) Y= -0.0867x^2+12.318x-405.74 2) Y= 0.0028x^2+0.2613x+56.976 where x is the rebound angle, and Y is the compressive strength (MPa). The results show a strong correlation between rebound angle and compressive strength, validating the effectiveness of the proposed method. The study demonstrates that this technique is economical, simple, and suitable for real-time field applications.

DOI: https://doi.org/10.5281/zenodo.19448127

Monitoring Of Progressive Concrete Compressive Strength Using NDT

Authors: Mr. Rohit S. Kamble, Dr.Ganesh A. Hinge, Mr. Mangesh D. Kevadkar

Abstract: This study investigates the monitoring of progressive compressive strength development in concrete using non-destructive testing (NDT) techniques, aiming to provide a reliable, in-situ alternative to traditional destructive testing. An experimental program was conducted on concrete specimens (mixes with varying strength ranges) to track strength evolution at early and mature ages, specifically 7, 28, and 90 days. The NDT techniques, primarily focusing on Ultrasonic Pulse Velocity (UPV) and the Rebound Hammer (RH) test, were utilized to estimate the compressive strength (SonReb method). The research confirms that while sole NDT methods possess inherent inaccuracies (roughly $pm-20% for RH), combining RH and UPV significantly improves the precision of in-situ concrete strength predictions.This study presents a systematic investigation of the monitoring of progressive compressive strength of M30 Concrete using a pendulum-based non-destructive testing (NDT) technique. The research is based on the principle of the coefficient of restitution, where the rebound angle of a pendulum after impact is correlated with the compressive strength of concrete. Traditional strength evaluation using the Compression Testing Machine (CTM) is destructive and unsuitable for in-situ monitoring. To overcome this limitation, a pendulum-based device is used to measure rebound angle at different curing ages (2, 7, 14, 21, and 28 days).Two empirical relationships were developed 1) Y= -0.0867x^2+12.318x-405.74 2) Y= 0.0028x^2+0.2613x+56.976 where x is the rebound angle, and Y is the compressive strength (MPa). The results show a strong correlation between rebound angle and compressive strength, validating the effectiveness of the proposed method. The study demonstrates that this technique is economical, simple, and suitable for real-time field applications.

Optimization-Enhanced Data Mining For Plant Disease Detection: A Framework For Evaluating Complex Attribute Interactions And Unobserved Pattern Recognition

Authors: Swapnil Wagh, Ruchi Sharma, Ankit Temurnikar

Abstract: In the field of contemporary agriculture, the challenge of early recognition of plant diseases is one of the most vital occurrence since the possibility of consequential factors directly influences the harvest, food security, and economical continueance. Although data mining method has demonstrated much promise in identifying useful patterns in agricultural data sets, given its tendency to enumerate complicated interaction of attributes and the ability to notice patterns that are not observed or visible, it frequently has difficulty distinguishing the core. To handle the shortcomings, this research paper suggests an Optimization-Enhanced Data Mining Framework which combines modern data optimization algorithms with conventional methods of data mining in detection of the plant diseases. The systematically preprocesses the agriculture data, includes attribute and symptom variations, and makes use of optimization that is used to find out non-linear and combinatorial impacts that are not usually apparent to the conventional mining strategies. The comparison of all the parametric conditions (a change of the attributes, the effects of a combination of several methods, or the effects of a combination of several modulations) through the means of an experiment proved that optimization integration may contribute significantly to an increase in the disease detection accuracy, device strength, and prolonged computations. Findings have also shown a significant change in false detection trends and improvement in identifying indicators of subtle diseases, thus making the structure applicable in practice and as a component of agricultural decision-making systems. This study also bridges the gap between pattern mining and optimization-based learning in making the study feel scalable and versatile to vulnerable applications by the intelligent detection of plant disease. The results emphasize impact of preciseness data mining that is optimization-enhanced to enhance agricultural tools in managing the disease as well as disease diagnosis activities and sustainable farming in serving purpose of precision agriculture.

DOI: https://doi.org/10.5281/zenodo.19448564

A Review Of Cryptographic Solutions And Forensic Readiness In IoT And Network Security

Authors: Muhammad Ahmad, Hua Zhou, Tanzeela Bibi, Haider Ali

Abstract: In today's digital environment, the swift advancement of interconnected technologies has raised significant worries about data safety, privacy, and reliability. The Internet of Things (IoT), networking systems, and cloud services produce and transfer large quantities of sensitive information, leaving them susceptible to cyber threats and other security risks. This research offers a detailed evaluation of how cryptography, network protection, and digital forensics work together, highlighting their combined impact on securing communication, safeguarding data integrity, and ensuring effective investigation methods. The approach to research relies on a thorough examination and combination of available literature, with a focus on major developments in cryptographic methods, network defense strategies, and forensic analysis frameworks. Particular focus is given to Homomorphic Encryption (HE), which allows processing to occur directly on encrypted information without the need for decryption, thus increasing privacy in unreliable settings such as cloud services and IoT environments. Moreover, the research includes new strategies in blockchain-centered forensics, featuring automated cost management that aligns with regulations, mapping wallet interactions, and utilizing non-fungible tokens (NFTs) as reliable audit references to enhance transparency and responsibility. The results show that cryptographic methods ensure safe data transfer, while network security strategies defend systems against unauthorized access, misuse, and cyber intrusions. At the same time, digital forensics offers a scientifically supported method for finding, preserving, and examining digital proof, tackling key evidentiary issues in today’s cyber landscape. The integration of blockchain forensics and NFTs further boosts auditability, traceability, and trust, especially within decentralized finance (DeFi) setups and intricate digital transactions. In summary, the alignment of cryptography, network protection, and digital forensics creates a strong and forward-thinking security framework that improves data safety, helps with regulatory adherence, and enhances the overall durability of contemporary digital systems.

DOI: https://doi.org/10.5281/zenodo.19449141

Impact Of Agricultural AI Technologies On The Habitat And Feeding Patterns Of The Nilgai (Boselaphus Tragocamelus) In Hardoi District: A Comprehensive Review

Authors: Kiran Lata Verma

Abstract: Scientifically, the Nilgai is known as the largest antelope in Asia, which is scientifically referred to as Boselaphus tragocamelus. Such beautiful animals can be found in northern states of India, and, where particularly prevalent, are especially popular in rural areas such as the Hardoi district of Uttar Pradesh. The environment and development of agriculture has changed dramatically over the past few years, in large part because of farming improvements, technology and AI here. These innovations have transformed the agricultural practices and surrounding habitat for the Nilgai. Yet, as farming practices change, new hurdles arise for these lovely animals. Their patterns of doing things are changing, their routes of travel are shifting, and they are discovering food in various manners all inspired by the new kinds of things that are occurring in their habitats. We will discuss recent trends of AI in agriculture drone surveillance, precision farming sensors, automated irrigation, GPS-guided tractors, and AI-based crop protection that are redefining the ecology of the Nilgai. We would like to explore the impact of contemporary farming to wildlife and its larger consequences for these stunning antelopes. In this review, we extract and discuss evidence from diverse ecological studies and reports, including those from the Hardoi region, zoological studies, and interdisciplinary environmental research, with the aim to evaluate the effects of AI-powered agriculture on the habitat and behavior of Nilgai, a large antelope species. We discuss how these advances in technology could influence feeding behaviors, migration patterns, risk perceptions, and interactions with human beings. While AI brings the potential of enhancing throughput and reducing labor costs in agriculture, it poses a series of unexpected ecological problems. These elements include habitat fragmentation, changes in the provision of food over space and time, a reduction in natural food sources, changing wildlife behavior, and escalating conflicts between humans and Nilgai. These interlinks reveal that on the one hand there are great pros and cons to implementing modern farming processes on wildlife and the places they live. The paper concludes that sustainable coexistence will involve the design of AI-enhanced mitigation frameworks, wildlife-friendly precision farming, predictive conflict-monitoring models, and landscape-level ecological planning. Such interfacing of zoology, AI modelling, conservation policy, and traditional ecological knowledge is required to avoid the danger of advancing technologically at the expense of Hardoi’s ecological stability and Nilgai survival.

DOI: https://doi.org/10.5281/zenodo.19452615

Experimental Study On Strengthening Of Rc Beam Using Glass Fibre Reinforced Polymer Sheets

Authors: Mr. Om Shrikant Zambare, Prof. V. S. Dhote

Abstract: Many concrete structures around the world suffer from corrosion, lack of detail, lack of connections between lines and lines, additional services, etc. It now needs repair, maintenance, or reconstruction due to various reasons such as. FRP composites are considered an effective method for repairing and strengthening RCC structures. This paper reports the influence of GFRP sheets wrapped around RCC beams for strengthening. Beams of size were cast and tested. Some of them are without GFRPS, and the remaining beams wrapped in different lay-up patterns with 90 and 45 degrees of inclination of GFRP sheets were subjected to shear and flexural tests. Initial crack load and ultimate failure load have been observed and noted. Experimental results indicate a significant increase in the initial and ultimate load-carrying capacity of GFRPS wrapped beams compared to unwrapped beams.

DOI:

Handwritten Recognition System Using OCR

Authors: Dr. A. V. Mane, Anushree Kalloli, Ishwari Kamble, Amol Kote, Tejas Padwal

Abstract: This project presents an Image Text Recognition and Translation System that extracts text from images and converts it into editable and translatable digital content. The system uses image processing techniques to enhance image quality and improve text detection accuracy. By integrating Tesseract OCR, the application efficiently recognizes printed and partially handwritten text from images. After extraction, the recognized text is translated into different languages using an integrated translation module, making the system useful for multilingual communication. Additionally, the system stores the original and translated text in a database, enabling users to maintain a history of their data for future reference. This project aims to reduce manual effort, improve productivity, and provide a user-friendly solution for text extraction and translation. It can be applied in areas such as document digitization, education, and travel assistance. Future improvements may include enhanced handwriting recognition, voice output, and mobile application support

DOI: https://doi.org/10.5281/zenodo.19458559

CIS: CYBER IMMUNITY SCORE

Authors: Prof. Bhoge.S.S, Rohan Pawar, Kedar Tible, Bhargav Fanse, Suchit Navsare

Abstract: The Customer Interaction System (CIS) is a web-based application developed to enhance the interaction between users and service providers through a centralized digital platform. The primary objective of the system is to simplify user communication, manage service-related activities efficiently, and provide a seamless user experience through an intuitive interface. The system allows users to register, log in, and access a personalized dashboard where they can explore available services, submit requests, and track their interactions. It also includes an administrative module that enables administrators to manage user data, monitor system activities, and handle service requests effectively. The application is developed using modern web technologies such as HTML, CSS, and JavaScript for the front-end interface, along with backend integration for data processing and storage. The system ensures proper data handling, user authentication, and structured information management to maintain reliability and performance. By digitizing the interaction process, the Customer Interaction System reduces manual efforts, improves communication efficiency, and enhances overall system transparency. This project demonstrates the practical implementation of web development concepts and provides a scalable solution that can be further extended with advanced features in the future.

DOI: https://doi.org/10.5281/zenodo.19466741

Medishield IDS- Protecting Medical WSNs Like A Shield

Authors: Aysha Salim, Athul S, Abdul Rahman, Alex Abraham

Abstract: MediShield IDS is an IoT-based smart health mon- itoring and security system designed to provide continuous, secure, and intelligent patient care in hospitals. The system includes a wearable smart health band that monitors vital signs like heart rate, body temperature, oxygen saturation (SpO), and patient movement in real time. The band uses biomedical sensors and an ESP32 microcontroller for efficient data collection and wireless communication. The collected health data is securely sent to a cloud-based platform or mobile app, allowing doctors and caregivers to monitor patient health remotely and respond quickly to critical situations. Besides health monitoring, the system provides indoor navigation and real-time patient tracking with Bluetooth Low Energy (BLE) beacons and motion analysis using an accelerometer. This feature helps elderly and disabled patients navigate hospital areas safely with turn-by-turn guidance through vibrations or display alerts on the wearable device. At the same time, healthcare staff can track patient locations via a web-based interface, improving emergency response, patient management, and operational effectiveness. To tackle growing cy- bersecurity issues in medical Wireless Sensor Networks (WSNs), MediShield IDS includes strong security measures like end-to-end data encryption, device authentication, and anomaly detection to block cyberattacks, data spoofing, and unauthorized access. By combining health monitoring, indoor navigation, real-time tracking, and intrusion detection into one wearable solution, the system improves patient safety, privacy, independence, and hos- pital efficiency. MediShield IDS offers a secure smart healthcare solution that meets the needs of modern digital hospitals.

DOI: https://doi.org/10.5281/zenodo.19469605

DeepShield: A Hybrid Deep Learning Framework For Real-Time Deepfake Detection Using Spatial And Temporal Cues

Authors: Rajnandini Birajadar, Sanika Shinde, Krutika Sane, Shivani Khopkar

Abstract: Deep learning has demonstrated remarkable success in solving complex problems across various domains, such as big data analytics, computer vision, and human-level control. However, the same advancements in deep learning have also given rise to applications that pose threats to privacy, democracy, and national security. One such application is deepfake technology, which leverages deep learning algorithms to create convincingly realistic fake images and videos that are indistinguishable from authentic ones. Consequently, the need for technologies capable of automatically detecting and assessing the integrity of digital visual media has become imperative. This paper aims to present a comprehensive survey of the algorithms employed to create deepfakes and, more importantly, the methods proposed in the literature for detecting deep fakes. The survey delves into extensive discussions on the challenges, research trends, and future directions concerning deepfake technologies. By reviewing the background of deepfakes and examining state-of the-art deepfake detection methods, this study provides an inclusive overview of deepfake techniques, thereby facilitating the development of novel and robust methods to combat the increasingly sophisticated deep fake threats in conclusion, this survey paper provides a comprehensive overview of deepfake techniques and detection methods. By synthesizing the existing literature and highlighting research trends and challenges, it aims to support the development of novel and effective approaches to combat the growing threat of deep fakes, ensuring the integrity, privacy, and security of digital visual media in an increasingly complex and interconnected world.

AI-Acoustic Guard: Real-Time Human Fall Detection Using TinyML And Edge Computing

Authors: Mrs K G Suhirdham, Srinithi A, Vishalini S, Yuvasri S

Abstract: Falls are a leading cause of injury, particularly among elderly individuals and vulnerable populations, necessitating efficient and real-time monitoring solutions. This paper presents AI-Acoustic Guard, a novel human fall detection system leveraging TinyML and edge computing for low-latency, privacy-preserving operation. Unlike conventional vision-based approaches, the proposed system utilizes acoustic signals captured through low-power microphones to identify characteristic sound patterns associated with human falls. A lightweight deep learning model is trained and optimized using TinyML techniques, enabling deployment on resource-constrained edge devices such as microcontrollers. The system processes audio data locally, eliminating the need for continuous cloud connectivity and ensuring data security. Feature extraction methods, including Mel-frequency cepstral coefficients (MFCCs), are employed to enhance classification accuracy. Experimental results demonstrate high detection accuracy, low false alarm rates, and minimal power consumption. The proposed approach is cost-effective, scalable, and suitable for real-world applications such as smart homes, assisted living facilities, and healthcare monitoring systems, offering a reliable solution for timely fall detection and emergency response.

DOI: https://doi.org/10.5281/zenodo.19471348

Post-TB Lung Risk Forecaster: A Real-Time Clinical Decision Support System.

Authors: Mrs.S.Subha , Abinaya M, Dharshini S , Hari priya J

Abstract: Diabetic retinopathy (DR) is a leading cause of vision impairment among individuals with diabetes, making early detection and timely treatment essential. This study presents an automated approach for diagnosing diabetic retinopathy using digital image processing techniques applied to retinal fundus images. The proposed system enhances image quality through preprocessing steps such as noise reduction, contrast enhancement, and normalization. Key pathological features, including microaneurysms, hemorrhages, and exudates, are detected using segmentation and morphological operations. Feature extraction techniques are employed to quantify these abnormalities, followed by classification using machine learning algorithms to determine the severity of DR. The model is evaluated using standard retinal image datasets, demonstrating improved accuracy, sensitivity, and specificity compared to traditional manual screening methods. This approach reduces dependency on expert ophthalmologists and enables scalable, cost-effective screening, particularly in resource-limited settings. The results indicate that digital image processing combined with intelligent classification can significantly enhance early diagnosis and management of diabetic retinopathy, ultimately preventing vision loss.

DOI: https://doi.org/10.5281/zenodo.19471429

CareerMate AI – GenAI Based Career Guidance And Alumni Mentorship Platform

Authors: Dr.M.Suganthi, Hansika P, Jasvanthika S, Joice Merlin X

Abstract: CareerMate AI is a Generative AI–based career guidance and alumni mentorship platform designed to support students and early professionals in making informed career decisions. The system leverages advanced natural language processing and machine learning techniques to provide personalized career recommendations based on user interests, skills, academic background, and market trends. By integrating a dynamic knowledge base with real-time labor market insights, the platform offers tailored guidance on career paths, required competencies, and learning resources. A key feature of CareerMate AI is its alumni mentorship module, which connects users with experienced professionals for guidance, networking, and real-world insights. The platform uses intelligent matching algorithms to pair users with suitable mentors, enhancing the quality of interactions. Additionally, conversational AI enables users to engage in interactive, human-like discussions, making the guidance process more accessible and engaging. Continuous learning mechanisms allow the system to refine recommendations over time. Overall, CareerMate AI bridges the gap between education and industry by combining AI-driven insights with human mentorship.

DOI: https://doi.org/10.5281/zenodo.19471479

E-Auction Portal For Seized Vehicles

Authors: K.G.Suhirtham, M.Rohith, S. Mugesh, V.Muthuprasanna

Abstract: An E-Auction Portal for Seized Vehicles is an online system developed to simplify and modernize the auctioning process of vehicles confiscated by authorities such as police departments, banks, or financial organizations. Conventional auction methods are often inefficient, lack transparency, and involve manual procedures that may lead to delays and unfair practices. The proposed portal utilizes web-based technologies to create a secure, transparent, and accessible platform for conducting auctions digitally. Authorized agencies can upload detailed information about seized vehicles, while registered users can conveniently participate in bidding from any location. The system supports real-time bidding, automated bid management, and secure payment processing to ensure fairness and reliability. It also maintains detailed records of all transactions, enhancing accountability and minimizing the risk of fraud. By increasing accessibility and participation, the platform helps achieve better price realization. Overall, the system provides an efficient, transparent, and dependable solution for managing seized vehicle auctions.

DOI: https://doi.org/10.5281/zenodo.19471546

Diabetic Retinopathy Diagnosis Using Digital Image Processing

Authors: Mrs.K.M.Swarna Devi, Dharshini M, Dhanusha sri K, Kalaivani S

Abstract: Diabetic retinopathy (DR) is a leading cause of vision impairment among individuals with diabetes, making early detection and timely treatment essential. This study presents an automated approach for diagnosing diabetic retinopathy using digital image processing techniques applied to retinal fundus images. The proposed system enhances image quality through preprocessing steps such as noise reduction, contrast enhancement, and normalization. Key pathological features, including microaneurysms, hemorrhages, and exudates, are detected using segmentation and morphological operations. Feature extraction techniques are employed to quantify these abnormalities, followed by classification using machine learning algorithms to determine the severity of DR. The model is evaluated using standard retinal image datasets, demonstrating improved accuracy, sensitivity, and specificity compared to traditional manual screening methods. This approach reduces dependency on expert ophthalmologists and enables scalable, cost-effective screening, particularly in resource-limited settings. The results indicate that digital image processing combined with intelligent classification can significantly enhance early diagnosis and management of diabetic retinopathy, ultimately preventing vision loss.

DOI: https://doi.org/10.5281/zenodo.19475022

Design And Development Of A Smart Temperature Controlled Sustainable Packaging System From Agriculture Waste

Authors: J.H.Ashwin, S.Gokul, M.Vigneshwaran, Dr.T.Sengolrajan

Abstract: The design and development of a smart temperature-controlled sustainable packaging system fabricated from agricultural waste materials. Banana fiber, sugarcane bagasse, rice husk and coconut coir are utilized to form a biodegradable composite structure bonded using a starch-based adhesive. Phase change material (PCM) is integrated to provide passive thermal energy storage and temperature stabilization during transit. An embedded smart control unit comprising an ESP32 microcontroller and a DS18B20 temperature sensor enables continuous monitoring of internal temperature. A E27 Infrared heat bulb is automatically activated when the temperature falls below a predefined threshold, ensuring controlled thermal regulation. The proposed system offers an eco-friendly, cost-effective and energy-efficient solution for transporting temperature-sensitive goods.

DOI: https://doi.org/10.5281/zenodo.19475939

To Study The Biological Carbon Sequestration Using BHOPAL Lakes Microalgae

Authors: Deeksha vijayvargiya

Abstract: Microalgae represent efficient photosynthetic microorganisms with the potential to serve as natural air purifiers by generating oxygen and reducing CO2 levels. Microalgae can play an important role in reducing carbon dioxide (CO₂) levels in cities. They grow fast, use sunlight, and absorb CO₂ during photosynthesis. When used in small urban spaces—like near traffic signals, on building walls, and in decorative fountains—microalgae can clean the air and improve the environment. This paper explains how microalgae can be used in such city setups, their advantages, and their challenges. It also discusses real and possible examples of building façades, fountains, and traffic junctions using algae-based systems to trap CO₂ and create cleaner air.

DOI:

INDRA: Intelligent Networked Document Retrieval Agent With Agentic AI For Autonomous Institutional Decision Support

Authors: Arya Sakore, Mansi Wagh, Manjeet Singh, Kavita Sharma, Vidya Dhoke, Manjusha Tatiya

Abstract: Educational institutions often face challenges in providing timely, accurate, and personalized academic assistance due to the absence of centralized and intelligent support systems. This paper presents INDRA, an integrated agentic artificial intelligence–based academic assistance framework designed to enhance student learning experiences in technical education environments. The system leverages a multi-agent architecture combining large language models with task-specific modules to deliver functionalities such as syllabus-aware tutoring, automated study planning, attendance tracking, and real-time academic analytics. By incorporating context-aware filtering based on department and semester, the system ensures that all generated responses remain relevant, structured, and aligned with institutional curricula. The framework utilizes a modular pipeline consisting of data processing, agent orchestration, and user interaction layers, enabling scalable and efficient deployment. Experimental evaluation demonstrates improvements in information accessibility, reduction in manual effort, and enhanced academic decision-making support. The system maintains low response latency while ensuring high usability and adaptability across different academic domains. Overall, the proposed approach provides a practical and intelligent solution for bridging the gap between students and academic resources, contributing to improved learning efficiency, engagement, and institutional productivity.

Transforming Public Service Delivery Through Digital HRM And E-Governance: A Contextualized Framework For Local Government Units In Northern Cebu

Authors: Soleste Kiamco del Mar

Abstract: The rapid evolution of digital technologies has transformed governance systems worldwide, prompting Local Government Units (LGUs) to adopt innovative approaches to public service delivery. This study examines the role of Digital Human Resource Management (DHRM) and e-governance in enhancing efficiency, transparency, and citizen satisfaction in selected LGUs in Northern Cebu, Philippines. Anchored on New Public Management (NPM), the Technology Acceptance Model (TAM), and the Diffusion of Innovations Theory, the research employed a descriptive-survey design involving 144 respondents comprising internal stakeholders (LGU officials and employees) and external stakeholders (citizens and community representatives). Quantitative data were analyzed using descriptive statistics, including frequency, percentage, and weighted mean, while qualitative data were subjected to thematic analysis. Findings reveal that LGUs demonstrate an “evident” level of digital maturity in HRM functions (M = 4.08) and e-governance services (M = 4.01). The integration of digital systems significantly improves public service delivery, with a “highly evident” impact on efficiency and transparency (M = 4.24). However, persistent challenges include budget constraints, infrastructure limitations, and insufficient training programs. The study concludes that digital transformation enhances governance outcomes when supported by institutional readiness, leadership commitment, and continuous capacity building. A contextualized framework is proposed to guide LGUs in optimizing digital transformation initiatives and sustaining citizen-centered governance.

DOI: https://doi.org/10.5281/zenodo.19479685

AI-Powered Adaptive Frequency Antenna For Seamless Wireless Communication

Authors: Mr. Sunil Kondiram Janrao, Mr. Hemraj Tukaram Gangurde, Miss. Jadhav Savita Yogiraj, Miss. Rupali Shravan Salmuthe, Dr. Prashant Thakare

Abstract: Wireless communication systems today demand high reliability and adaptability as signal quality constantly varies due to interference, environmental obstacles, and user mobility. However, conventional fixed-frequency antennas lack the ability to adjust to such dynamic conditions, resulting in signal degradation, reduced coverage, and unstable communication. To address this limitation, this project proposes an AI-Powered Adaptive Frequency Antenna System capable of intelligently switching between multiple microstrip patch antennas tuned to 2.4 GHz, 3.7 GHz, and 5.5 GHz. A Software Defined Radio (SDR) module is used to capture real-time signal data, while a Raspberry Pi processes key parameters such as Received Signal Strength Indicator (RSSI) and Signal-to-Noise Ratio (SNR) using AI-based decision-making algorithms. Based on the analysis, the system automatically selects and activates the antenna that provides the best communication performance through a relay or RF switch. This adaptive approach ensures stable connectivity, improved signal quality, optimized bandwidth usage, and energy-efficient operation. The proposed system offers a low-cost, scalable, and intelligent solution suitable for applications such as IoT networks, smart infrastructure, mobile communication, and next-generation wireless systems.

WordBlitz: Redefining Typing Through Real-Time Gamified Learning

Authors: Rajan Parmar, Aman Rauniyar, Vikas Chawla, Preety Chaudhary

Abstract: WordBlitz is a real-time multiplayer typing game aimed at improving typing skills through competitive gameplay. It uses the MERN stack with WebSocket integration, providing real-time interactions, dynamic leaderboards, and feedback on speed and accuracy. Players can join multiplayer sessions, race against others in real time, and track their performance on a global leaderboard. The game is developed using modern web technologies, including MongoDB, Express.js, React.js, and Node.js to ensure scalability, responsiveness, and a smooth user experience. The leaderboard is dynamically updated, providing users with immediate feedback on their performance and allowing them to compare their typing speeds with other players. WebSockets are used for real-time communication, enabling the synchronization of game state between players. Additionally, the game supports multiple languages and session selection so that users will get the best experience. In this game, we added a room join option so that using one ID, multiple players can join easily. Future updates will focus on adding more multiplayer features and optimizing the game engine.

Iot Based Vehicle Tracking And Monitoring System

Authors: M.Santhoshkumar, Dr.P.kavitha

Abstract: The rapid growth of urban transportation has created a pressing need for intelligent vehicle tracking and monitoring solutions. This project proposes an IoT-based system that integrates GPS modules, sensors, and cloud computing to enable real-time tracking and monitoring of vehicles. The system collects data such as location, speed, fuel consumption, and engine health, transmitting it to a centralized platform via wireless communication. Authorized users can access this information through mobile or web applications, ensuring transparency, safety, and efficiency in transportation management. The proposed solution enhances fleet management by providing route optimization, predictive maintenance, and driver behavior analysis, while also offering security features such as geofencing and unauthorized movement alerts. By leveraging IoT technology, the system contributes to smarter cities, improved logistics, and safer travel experiences. This project demonstrates how IoT can transform conventional transportation into an intelligent, data-driven ecosystem. The outcome is a scalable solution that supports sustainability, efficiency, and enhanced road safety.

DOI: https://doi.org/10.5281/zenodo.19484221

AI Tool/mobile App For Indian Sign Language(ISL) Generator From Audio-visual Content In English/Hindi To ISL Content And Vice-versa

Authors: Dr. Harsha R. Vyawahare, Sukhada Shripad Tare, Ashwini Nitin Shingane, Shreya Sunil Shinde, Bhavika Suraj Jain

Abstract: This paper presents a lightweight and practical bidirectional communication system designed to translate between speech and Indian Sign Language (ISL) using machine learning and computer vision techniques. The system operates in two modes: (i) speech-to-ISL translation, where spoken input is converted into text and further mapped into a sequence of ISL alphabet images, and (ii) ISL-to-text and speech translation, where hand gestures captured through a webcam are recognized using a Convolutional Neural Network (CNN) model and converted into readable text and audio output. Unlike existing approaches that rely on complex natural language processing techniques or computationally expensive 3D avatar rendering, the proposed system focuses on simplicity, real-time performance, and ease of implementation. By utilizing a TensorFlow/Keras-based CNN model for gesture recognition and a predefined ISL image dataset for visual representation, the system achieves efficient and accurate translation with low computational requirements. The system is implemented using Python with libraries such as OpenCV, Streamlit, speech_recognition, and pyttsx3, enabling an interactive and user-friendly interface. The proposed solution provides a cost-effective and accessible tool to bridge the communication gap between hearing individuals and the Deaf and Hard-of-Hearing (DHH) community, making it suitable for real-world applications.

DOI: https://doi.org/10.5281/zenodo.19484295

A Structured Approach For Assessment And Attainment Of Program Outcomes Under NBA Criterion 3 Using An Excel-Based Evaluation Tool

Authors: Mr. Sourabh S. Lohar, Mr. Nilesh N. Kokare

Abstract: Outcome-Based Education (OBE) has become an essential framework for engineering education under the National Board of Accreditation (NBA). Criterion 3 of NBA focuses on the evaluation and attainment of Program Outcomes (POs) through measurable and transparent processes. This paper presents a structured and practical methodology for calculating CO and PO attainment using an Excel-based automated tool. The methodology integrates both direct and indirect assessment techniques and provides a systematic approach for data collection, analysis, and interpretation. A case study from a Mechanical Engineering course is included to demonstrate real implementation.

DOI:

A Study To Evaluate The Effectiveness Of Video Assisted Teaching On Knowledge Regarding Online Game Addiction Among School Children At Selected School, Kannur District

Authors: Prof. Dr. Usha V, Prof. Dr. J Sathya Shenbega Priya, Ms. Amritha. S. Nair, Ms. Anaya. P. Sebastian, Mr. Anjaneyaraj. T. K, Mr. Anurag. K. K, Mr. Anugrah. K, Ms. Arathy Saju, Ms. Ashlin Anil, Ms. Aswathy. T. Biju

Abstract: The present study was conducted to evaluate the effectiveness of video-assisted teaching on knowledge regarding online game addiction among school children aged 11–13 years in a selected school at Kannur district. Excessive involvement in online games has been shown to negatively impact physical, emotional, and academic performance. A quantitative research approach with a pre-experimental one group pretest post-test design was used. 30 students from classes VI and VII were selected via convenience sampling. Data were collected using a semi structured questionnaire before and after a 13minute video intervention. Pre-test results showed 43.3% inadequate and 56.67% moderate knowledge. Post-test results significantly improved, with 86.67% achieving moderate knowledge levels. The study concluded that Video Assisted Teaching is an effective educational tool for enhancing awareness.

Sediment Transport Analysis On Mandavya River Using Hec-Ras Software

Authors: Dr. D. Gouse Peera, A.Nandha Kishor, B.Aparna, N.Mahesh Babu, C. Sri Ranga Hemanth, P. Amarnadh

Abstract: Sediment transport is an essential aspect of river engineering that affects channel stability, flood risks, and hydraulic structures. This study focuses on analyzing sediment behavior in the Mandavya River using HEC-RAS modelling software. The objective is to evaluate sediment movement, erosion, and deposition under different flow conditions by incorporating river geometry, flow data, and sediment characteristics into the model. Both steady and unsteady flow simulations were carried out, and various sediment transport functions were tested to determine their suitability. The analysis highlights sediment load distribution, bed level changes, and areas prone to aggradation and degradation. Results indicate that sediment transport is highly influenced by discharge variations and channel slope. The model successfully identifies critical zones of erosion and deposition, supporting effective river management and flood mitigation planning. Overall, the study demonstrates that HEC-RAS is a reliable tool and provides a strong scientific basis for sustainable river basin management and infrastructure development.

Early Detection Of Mental Distress In GENZ

Authors: Vanshika Thakur, Sushmita Guha, Nafiya Kausar N, Mrs.Jayahsree Kudari

Abstract: Mental distress among Generation Z has emerged as a significant public health concern due to increased exposure to digital environments, academic stress, and social pressures. This survey paper analyzes existing methodologies for early detection of mental distress using machine learning and artificial intelligence techniques. The study reviews various approaches including Natural Language Processing (NLP), sentiment analysis, and deep learning models applied to social media and behavioral datasets. It also explores commonly used datasets, particularly from Kaggle, and evaluates their effectiveness in predictive modeling. The survey identifies limitations in current systems such as data bias, privacy concerns, and lack of real-time adaptability. Furthermore, research gaps and future directions are discussed, emphasizing the need for multimodal data integration and ethical AI deployment.

A Quantitative Study to Assess Whether Pollutants Influence Immune Diseases (Asthma). A Case Study of Mufulira Clinic 1 Urban Health Centre

Authors: Magret Nkandu

Abstract: This quantitative study examined the impact of environmental pollutants on immune diseases, specifically asthma, at Mufulira Clinic 1 Urban Health Centre in Zambia. The research addressed a key public health issue due to the increasing prevalence of asthma in urban areas, worsened by exposure to pollutants like Sulphur Dioxide (SO₂), Carbon Monoxide (CO), and cigarette smoke. The main goals were to explore the link between Sulphur Dioxide (SO₂) exposure and asthma symptoms, study the connection between Carbon Monoxide (CO) exposure and asthma severity, and assess how cigarette smoke exposure influences the development of asthma among patients attending the clinic. The study focused on a clearly defined group of asthma patients who had been receiving treatment for at least six months, ensuring the results were relevant to those most affected. Limitations included a relatively small sample size of 87 participants, which may have affected the ability to generalize the findings. Additionally, relying on self-reported exposure data introduced potential bias and inaccuracies in measuring true pollutant exposure levels. Delimitations were set to keep the scope manageable, with the study concentrating only on patients from Mufulira Clinic 1 and excluding individuals from nearby clinics or rural areas. The results revealed significant links between pollutant exposure and the severity of asthma symptoms, offering valuable insights into how environmental factors influence respiratory health. By identifying specific pollutants that worsen asthma, the research aims to guide public health strategies and interventions to reduce asthma rates in urban populations. Overall, the study emphasized the urgent need for targeted interventions and policy changes to improve air quality and address the increasing asthma burden in urban Zambia.

DOI: http://doi.org/

Planning, Design Analysis in Staad.Pro and Detailed Estimation with Valuation of a Residential Building (G+2)

Authors: R. Kiruthika

Abstract: The construction of residential buildings requires proper planning, structural safety, and cost efficiency. This study presents the planning, structural analysis, estimation, and valuation of a Ground plus Two (G+2) residential building. The building is designed to ensure efficient space utilization, proper ventilation, and user comfort. Structural analysis is carried out using STAAD.Pro to evaluate the behavior of the building under different loading conditions such as dead load and live load. The software-based approach provides accurate results for bending moments, shear forces, and deflections. Based on the analysis, structural members are designed to ensure safety and stability. A detailed estimation is performed to calculate the quantity of materials and overall cost of construction. Finally, valuation of the building is carried out considering construction cost and depreciation. The study demonstrates that integrating planning, analysis, and estimation results in an economical and structurally sound residential building.

Development Of Power Generation From Waste Heat In Industries Using Thermoelectric Generator

Authors: Dr. N. N. Khobragade, Shivam S. Sawargave, Vaishnavi S. Karhale

Abstract: The increasingly worldwide problem regarding rapid economy development and a relative shortage of energy, the internal combustion engine exhaust waste heat and environmental pollution has been more emphasized heavily recently. Out of the total heat supplied to the engine in the form of fuel, approximately, 30 to 40% is converted into useful mechanical work. The remaining heat is expelled to the environment through exhaust gases and engine cooling systems, resulting in to entropy rise and serious environmental pollution, so it is required to utilized waste heat into useful work. As waste heat recovering techniques, such as thermoelectric generator (TEG) is developed, Due to distinct benefits of thermoelectric generators, they have become a promising alternative green technology.

DOI:

Real-Time Defect Detection Using Edge AI In Smart Manufacturing Systems

Authors: Dr. Pankaj Malik, Vinayak Oberoi, Maanya Bhatia, Akshat Ghatewal, Tanishq Garg

Abstract: Ensuring zero-defect production in smart manufacturing demands fast, accurate, and intelligent inspection systems. However, conventional cloud-based defect detection approaches struggle with high latency, excessive bandwidth usage, and delayed decision-making, limiting their effectiveness in real-time industrial environments. To overcome these challenges, this paper proposes a novel Edge AI-driven framework for real-time defect detection, where optimized deep learning models are deployed directly on edge devices for instant analysis at the production line. The proposed system seamlessly integrates industrial vision sensors, edge computing units, and cloud platforms to achieve scalable, efficient, and intelligent quality control. Extensive experiments conducted on the MVTec AD dataset demonstrate that the proposed model achieves a high detection accuracy of 96%, while maintaining an ultra-low inference latency of 20 ms, significantly outperforming traditional cloud-based systems with latency exceeding 150 ms. Furthermore, the framework reduces bandwidth consumption by approximately 60%, enabling faster response times and efficient resource utilization. These results highlight the effectiveness of the proposed approach in delivering low-latency, high-accuracy, and scalable defect detection, making it a promising solution for next-generation Industry 4.0 manufacturing systems.

DOI: https://doi.org/10.5281/zenodo.19509358

AI Based Online Complaint Management System

Authors: Mr. M Anand, Balamurugan M, G Murali Krishna, Nakkala Lokesh

Abstract: In the contemporary era of digital transformation, efficient and automated complaint management systems have emerged as essential components of organizational service frameworks. This paper presents a comprehensive web-based online complaint management system developed using the Flask framework and database technologies in Python. The proposed system leverages modern web development techniques to perform real-time registration, tracking, and resolution of user complaints with high reliability. By integrating structured data storage models for complaint categorization and processing, the system enables efficient grievance handling that eliminates reliance on traditional manual-based mechanisms. The application provides an intuitive web interface that facilitates user complaint submission and administrative verification through structured input forms. Experimental results demonstrate that the system achieves processing efficiency exceeding 95% under optimal conditions, with consistent performance across varying operational scenarios. The lightweight architecture enables straightforward deployment on standard web servers, making the solution viable for diverse applications including service monitoring, issue tracking systems, and grievance redressal platforms in both academic and enterprise environments.

DOI: https://doi.org/10.5281/zenodo.19510108

Novel Comparative Analysis of Hybrid Solar–Wind Powered UPFC with Advanced Control Strategies for Power Quality Enhancement and Voltage Stability Improvement in Distribution Systems

Authors: U.Surekha, R. Priyadharshini, P.Deepa

Abstract: FACTS are the power electronic-based systems that provide fast and effective control of fundamental transmission characteristics like voltage, impedance, and phase angle. Hence allowing power networks to work in a far more efficient manner. The various FACTS devices directly integrated into gearbox lines improve power flow and system stability and include shunt and series compensators. One such device is the UPFC that provides full control capability by injecting the required voltage into the line through converter- based operation. Its dynamic responsiveness to voltage sags, swells, and harmonic disturbances is further refined through sophisticated control techniques such as neural- network-based modulation. The proposed approach to power quality improvement in a 230V distribution system is validated through simulation studies in MATLAB/Simulink and matching hardware implementation using renewable energy sources.

DOI:

Financial Performance Analysis Using Camel Model: A Study Of Selected Private Banks In India

Authors: Biswa Mohan Jena, Rajaram Majhi, Swarnalata Nayak, Bijay Jha

Abstract: Economic development is heavily reliant on the allocation and optimal use of resources, as well as the operational effectiveness of the many sectors, of which the banking industry is a key player. The banking industry facilitates monetary policy and aids in the promotion of capital formation, innovation, and monetization. Carefully assessing and analyzing bank performance is essential to maintaining a sound financial system and a productive economy. Using a five-year timeframe from 2017-18 to 2021-22, the current study aims to assess the financial performance of a few private sector banks in India using the CAMEL model.

DOI: https://doi.org/10.5281/zenodo.19510729

Modernist Aesthetics And Existential Anxiety In Nissim Ezekiel’s Poetic Oeuvre

Authors: Samapti Banerjee

Abstract: Nissim Ezekiel is the pioneering figure of post-colonial Indian poetry. He is a versatile genius. He works as an editor, dramatist, critic of art and Literature etc. For his enormous contribution in Literature he is awarded both the Sahitya Akademi Award in 1983 and Padma Shri award in the year 1988."Latter-Day Psalms", is the poetry collection for which he is awarded the Sahitya Akademi Award and received the prestigious prize from G. K. Gokak. He belongs to the Bene Israel Jewish family. He comes from an educated and erudite family. His father is the professor of Botany of Wilson College and his mother is the Principal of her school. Nissim Ezikiel has studied English from Wilson College, Bombay and Philosophy from London. He is brought up with embracing the Maharastrian tradition and culture and has accepted the Marathi language wholeheartedly. He has served as the Head of the Department of English in Bombay University and as a guest faculty and visiting Professor in the Leeds University. Nissim Ezekiel is the prominently post- colonial Indian English poet. Not only the Indian culture but also his family tradition has impacted his writings. Nissim Ezekiel bridges between Romanticism and Modernism in Indian English poetry.

DOI: https://doi.org/10.5281/zenodo.19510867

Ai-Driven Crop Recommendation System

Authors: M Uday Kanth Reddy, Modem Kalpana, Gowni Ushasree, Dharshini M, Dr. K Sasi Kala Rani

Abstract: Due to fragmented advisory systems, limited data access, and language barriers, smallholder farmers frequently face difficulties when choosing crops. Decisions are further complicated by the need for sustainable resource use and climate variability. In order to provide accurate, context-aware recommendations that increase productivity and sustainability, recent AI-driven crop recommendation systems integrate soil, weather, yield, and market data. These systems integrate data insights with agronomic knowledge through machine learning, ensemble models, and hybrid rule-based techniques. Conventional crop selection techniques frequently rely on local customs or experience, which might not always produce the best outcomes. This study offers an AI-based crop recommendation system that assists farmers in selecting appropriate crops based on soil and climate conditions in order to solve this issue. To choose the best crop for cultivation, the system examines critical factors like temperature, humidity, rainfall, soil pH, nitrogen (N), phosphorus (P), and potassium (K). The system was created using machine learning techniques, and an agricultural dataset was used to train several models, such as Random Forest, Decision Tree, and Support Vector Machine (SVM). The dataset was cleaned and made ready for analysis prior to training. Accuracy and confusion matrix analysis were used to assess these models' performance, and the Random Forest model outperformed the other tested algorithms. A straightforward prediction interface was developed to make the system useful and user-friendly. Users can enter soil and environmental values and instantly receive crop recommendations. This system can help farmers make better farming decisions, increase crop productivity, and promote more sustainable and effective farming methods.

DOI: https://doi.org/10.5281/zenodo.19511053

SENIMSU: Decentralized Snowmelt Water Recovery Through Passive Melting, Biochar Filtration, And Smart Water Quality Monitoring

Authors: S. Aityk, B. Alibi

Abstract: Regions experiencing cold climates, such as Kazakhstan, experience seasonal water shortages; however, these areas have abundant, albeit typically polluted, amounts of frozen precipitation. Approximately 90 percent of river run-off in Kazakhstan occurs during the springtime when snow melts, however much of this run-off evaporates or is un-treated and thus many rural communities are left without adequate access to clean water during the most critical time of year due to low precipitation. The paper presents new technology called SENIMSU which is a cost effective, decentralized technology that utilizes passive solar melting combined with multi-layered biochar filtration and an Arduino based sensor network with machine learning (ML) capabilities for assessing the quality of the treated water in real-time. Biochar is a type of charcoal produced by heating organic materials in the absence of oxygen at high temperatures. The biochar utilized in this study was generated by pyrolyzing agricultural waste at a temperature range of 450° – 500° C. The surface area and pH of the generated biochar were determined. Optimization of filtration parameters including particle size (0.5–2 mm), depth (15–30 cm), and flow rate (1–2.5 L/h) was achieved via Response Surface Methodology with a Box-Behnken experimental design. Simulated polluted snowmelt experiments utilizing the optimized filtration parameters resulted in significant reductions in turbidity (>85–>90%) (NTU of 30 to <5 NTU), TDS (>45–>55%) (ppm of 250 to <120 ppm), and pH stabilization (pH range of 6.4–7.2 to 7.0–7.2) and therefore meet WHO drinking water standards. Additionally, machine learning regression models (R² > 0.85) enabled accurate prediction of filtration efficiencies allowing for real-time optimizations. The estimated cost of the SENIMSU system per household unit will be compared to the 0+ estimated cost for commercial technologies and the use of locally generated biochar will eliminate the need to replace filters. SENIMSU is the first ML-integrated snow treatment system designed specifically for the climate conditions found in Central Asia and directly addresses SDG 6.1 (universal access to safe drinking water) and 6.3 (water quality improvement). Each unit has the capability to produce between 20 and 60 liters of clean water per day for a family of four to six people during the most critical time of the year when they require it (snow-melt period). Preliminary field testing conducted in the rural Akmola region of Kazakhstan reported that 92% of users accepted the use of SENIMSU systems and successfully operated them in temperatures below -15°C. FTIR analysis verified the adsorption mechanisms, whereas the decentralized nature of SENIMSU enables its replication across the 2.5 million rural Kazakhs who currently lack dependable access to clean water. This provides a scalable and locally sustainable means for communities located in cold climates around the world to address similar water security issues.

DOI: https://doi.org/10.5281/zenodo.19511933

A Review on Artificial Intelligence in Pharmaceutical Science

Authors: Krushna S. Bhutekar, Associate Professor Kishor B. Charhate, Khushi V. Kayande, Kiran. B. Nagre, Dr. Prafulla R Tathe

Abstract: Artificial Intelligence (AI) has emerged as a transformative technology in pharmaceutical science, revolutionizing various aspects of drug discovery, development, manufacturing, and healthcare delivery. This review highlights the current applications, advancements, and future potential of AI in the pharmaceutical sector. AI techniques such as machine learning, deep learning, and neural networks enable the efficient analysis of large and complex datasets, thereby accelerating drug discovery, target identification, and optimization of lead compounds. In pharmaceutical manufacturing, AI enhances process efficiency, quality control, and predictive maintenance through advanced automation and data-driven decision-making. Additionally, AI plays a crucial role in clinical trials, pharmacovigilance, personalized medicine, and drug delivery systems, including the development of novel nanocarriers.

DOI: http://doi.org/

An Analytical Study on Perception and Acceptance of Robotic-Assisted Surgical Interventions

Authors: Associate Professor Dr. Sushma Sumant, Ms. Rutuja Kiran Howal

Abstract: Surgical robotics has evolved into a lucrative alternative to conventional surgery. RAS, or robotic-assisted surgery, boasts many merits, such as precision, decreased invasiveness, and quicker recovery. The popularity of RAS, and its subsequent expansion, requires quality technologies on one hand; and the client society accepting the technology in surgical procedures on the other. This paper has reviewed the literature for public perception, as well as carried out a formal survey to collect public perceptions among 51 individuals spread across demographic areas. The prime manifestation of the study was that the public was fairly aware of RAS, yet lacked detailed knowledge of its application. For instance, participants perceived RAS to be safer and more precise than conventional surgeries, but raised concerns such as cost, utilitarianism (meaning the machine can malfunction or the software can fail), and loss of human control in their care. This study, therefore, rather highlights the need for the development of public education and engagement initiatives to build RAS technologies on a foundation of trust and acceptance. Robotic-assisted surgery is often approached with caution by patients especially when such patients have not had exposure to such a system or understand what they do fear of machine error absence of human control and the expense of the technology all stand as obstacles to its acceptance even though it has been tested and proven the issue is especially critical in developing countries like India where technological literacy and access to healthcare could be very disparate amongst the demographics. Key findings reveal that while approximately 85% of respondents had heard of RAS, a pervasive misconception exists regarding the autonomous nature of robotic surgery, with many wrongly believing robots operate independently. This misunderstanding contributes to common concerns such as machine malfunction, high cost, and a perceived lack of surgeon involvement. Despite these reservations, respondents widely acknowledged significant benefits, including smaller incisions, less pain, and faster recovery, and expressed a general willingness to consider RAS if recommended by a medical professional.

DOI: http://doi.org/

Complaint & Grievance Redressal System with Duplicate Detection and Smart Escalation

Authors: Mr. S. Krishna Reddy, P.Harika, N. Uma Tejaswini, J. Veera Shiva Kumar, N.Kevin

Abstract: The Complaint Grievance Redressal System (CGRS) was developed to improve the traditional method of managing student complaints by incorporating artificial intelligence for smarter complaint handling. In many institutions, the existing complaint process is manual, time-consuming, and difficult to track. To overcome these issues, the proposed system introduces semantic duplicate detection, automatic complaint routing, and priority-based escalation. The backend runs on Node.js and Express, with MongoDB handling data storage. When a student submits a complaint, the system uses Google Gemini 1.5 Flash to compare it against existing complaints in the database. If the similarity score hits 80 or above, the new submission is logged as a duplicate and the original complaint's duplicate count goes up by one. Once that count reaches three, the complaint is automatically bumped to high priority so it gets attention faster. Routing is automatic too. Based on the department and complaint type the student selects, the complaint goes straight to the right person either the Head of Department or the Training and Placement Officer. Students also get email notifications whenever their complaint status changes or gets resolved, so they're not left guessing. Access is split across four roles: Student, HOD, TPO, and Admin. Authentication uses JWT tokens, and each role only sees and does what it's supposed to nothing more. This paper presents the overall system architecture, artificial intelligence integration methodology, duplicate detection algorithm, and system performance evaluation.

DOI: http://doi.org/

Attention-Driven Low Light Image Enhancement Using Lightweight CNN

Authors: Mr. T. Sreenivasu, L. Vishnu Vardhan, Sk.Sayyad Baji, P.Yesuratnam, CH.Bhaskar, V.Prasanna Kumar

Abstract: Low-light image enhancement is an important task in computer vision that aims to improve visibility and preserve critical details in images captured under poor lighting conditions. In the base paper, a CNN-based method using a simple encoder–decoder architecture is employed for image enhancement. Although this approach effectively increases overall brightness, it presents several limitations. The model treats all pixels equally without prioritizing important regions, lacks an attention mechanism to focus on semantically significant features such as faces or text, and struggles to generalize effectively across diverse real-world lighting conditions. To address these limitations, this work proposes an enhanced Low-Light Image Enhancement (LLIE) model that integrates the Convolutional Block Attention Module (CBAM) into the encoder–decoder network. CBAM introduces both Channel Attention, which prioritizes important feature channels, and Spatial Attention, which focuses enhancement on key regions of the image.By incorporating these attention mechanisms, the proposed model improves brightness and clarity while preserving important structural details and features. As a result, the enhanced system produces higher visual quality images and becomes more suitable for practical applications such as surveillance systems, digital photography, and mobile vision applications.

DOI: https://doi.org/10.5281/zenodo.19548431

An Intelligent IoT-Driven ATM Security Framework With YOLOv5-Based Object Recognition

Authors: Golam Moula, Tania Akter Setu, Most. Mahmuda Akter, Sotabdi Rani, Onika Rahman

Abstract: In the era of digitalization, advanced technology is employed to enhance the safety and security of ATM users, ensuring the integrity of banking operations. The Real-time Automatic ATM Booth Security System aims to provide an innovative and effective solution by integrating sophisticated monitoring, detection, and reaction capabilities. To mitigate different types of robberies, we propose a security system for ATMs that detects specific threat objects based on real-time image analysis using the YOLOv5 Object Detection Algorithm. This system leverages an Arduino-based embedded platform to process real-time data collected through sensors. The module is designed to recognize and classify potential weapons, including screwdrivers, scissors, hammers, and rods. Upon detecting an individual carrying a weapon, the system generates an instantaneous alarm signal and autonomously triggers the closure of the ATM booth door, preventing the suspect from escaping. Another key feature of the proposed system is the real-time notification to the control center of the nearest police station and the respective bank. This approach ensures enhanced monitoring and control, ultimately improving ATM security.

DOI: https://doi.org/10.5281/zenodo.19549202

Design And Implementation Of A Safe Driving System For Real-Time Driver Behavior Analysis And Hazard Alerting Using Low Cost.

Authors: Payal N. Rathi, Roshani D. Khomane, Samruddhi D. Dharyekar, Sanskruti D. Dharyekar

Abstract: Driver inattention is a major contributor to road accidents worldwide, with fatigue responsible for nearly 20–30% of incidents each year. This study presents SafeDrive Alert System, a camera-based system that monitors driver behavior by detecting signs such as eye closure, head movement, and loss of focus using computer vision techniques. The system utilizes methods like Eye Aspect Ratio and facial landmark analysis, and its performance is evaluated using real-world driving data, achieving approximately 95% detection accuracy. Designed as a cost- effective solution, it can be easily deployed in vehicles such as taxis and trucks, with potential extensions for connected vehicle systems and additional safety features.

DOI:

Feastify : Online Food Ordering Website

Authors: Prashant Kothari, Dhruv Jagtap, Gulshan Kumar Singh, Ayush Singh, Pranay Gunnewar

Abstract: This research presents the design and implementation of a full-stack web application named FEASTIFY, developed to simplify and digitize the food ordering process. Traditional ordering methods often lead to delays, miscommunication, and inefficient order handling, especially in busy environments such as restaurants, food courts, and institutional canteens. The proposed system addresses these challenges by providing a centralized and user-friendly platform for both customers and administrators. The application enables users to browse menus, explore categorized food items, and place orders efficiently through an intuitive interface. Restaurant owners and administrators can manage food items, categories, and order statuses in real time. The system also incorporates authentication and role-based authorization to ensure secure and controlled access for different types of users . The backend is developed using Spring MVC and Spring Boot, with Hibernate/JPA for database operations and MySQL for data storage. The frontend utilizes HTML, CSS, JavaScript, Tailwind CSS, and React to provide a responsive user experience. Additional features such as order tracking and secure payment integration further enhance usability. Overall, the system improves efficiency, reduces manual effort, and provides a scalable solution for modern food ordering requirements.

Fake News Detection Using BERT + Graph Neural Networks

Authors: PRANAV J, Dr.Deepak Kr. Sinha

Abstract: The rapid dissemination of misinformation and fake news through online platforms has become a major challenge, impacting public perception, social stability, and decision- making. Existing transformer-based models, such as BERT, demonstrate strong textual understanding but fail to capture the relational and propagation dynamics of news across social networks. Conversely, graph-based models effectively model relationships between users, posts, and comments but lack semantic comprehension of textual content. To overcome these limitations, Zhang et al. (2024) introduced GBCA (Graph-BERT Co- Attention), which integrates Graph Convolutional Networks (GCNs) with BERT through a co-attention mechanism. While GBCA achieves significant performance gains, it still exhibits several drawbacks — it models homogeneous graphs, ignores temporal propagation, and underperforms in noisy or sparse network environments. This research proposes an enhanced hybrid framework, Heterogeneous Temporal Graph- BERT (HTGBERT), that extends GBCA by introducing heterogeneous graph modelling, temporal learning, and contrastive pretraining for robust fake news detection. The proposed model encodes textual semantics using BERT, constructs a heterogeneous social graph incorporating posts, users, comments, and entities, and applies a Temporal Graph Neural Network (TGAT/TGN) to learn propagation dynamics over time. A cross-modal contrastive learning module is employed to align text and graph representations, improving generalization and robustness to sparse or noisy data. Experiments will be conducted on benchmark datasets including Fake Newsnet, Twitter15/16, and PHEME, comparing the proposed model against baselines such as BERT- only, GNN-only, and GBCA. Performance will be evaluated using accuracy, F1-score, AUC, and time-to-detection metrics under event-separated evaluation protocols to ensure realistic generalization. The proposed HTGBERT framework is expected to achieve earlier, more accurate, and explainable detection of fake news by integrating semantic, structural, and temporal dimensions of information dissemination. This research not only advances hybrid fake news detection techniques but also contributes a reproducible, explainable, and temporally aware framework for real-world misinformation mitigation.

DOI:

Flow Analysis Of Supersonic Nozzle For Rocket Propulsion System

Authors: ML. Jorlin, L. Angel, A. Gowsalya, M. Mekala

Abstract: A supersonic nozzle normally has a converging-diverging shape, in which high-pressure gases pass through the throat and expand in the diverging section to attain supersonic velocity and generate thrust. When the exit area of the nozzle is changed, the expansion of gases at the outlet also changes, which directly affects pressure, velocity, Mach number, temperature, and thrust performance. In this project, the basic nozzle profile is maintained and only the exit area is varied by creating three different nozzle models to study the influence of exit area on the flow characteristics of a rocket propulsion system. The main purpose of changing the exit area is to identify which nozzle configuration provides better gas expansion and improved propulsion performance. A proper exit area can increase exhaust velocity, reduce pressure losses, improve thrust efficiency, and provide stable supersonic flow at the nozzle outlet. All three models are designed in Creo using Ni-Co-Cr- alloy-_wrought alloy as the selected material, and the flow analysis is carried out in ANSYS Fluent through CFD simulation. By comparing the results of the three models, the most suitable exit area can be identified based on better pressure distribution, velocity, Mach number, and thrust behavior. Thus, this project provides a simple and effective method for optimizing supersonic nozzle design and improving the efficiency of rocket propulsion systems.

DOI: https://doi.org/10.5281/zenodo.19552111

Simulation of Various Angles of Attack of NACA 23012 Airfoil Using CFD

Authors: ML. Jorlin, P Aarthi, S Deva Dharshini, R Rajeshwari

Abstract: This study focuses on the numerical investigation of the NACA 23012 aerofoil to improve its aerodynamic efficiency across different angles of attack. The primary goal is to enhance the lifting characteristics of the wing, which plays a crucial role during take-off and landing phases of flight. By implementing 20% split flaps, the aerofoil can generate higher lift without increasing the overall wing size, thereby effectively reducing drag and maintaining fuel efficiency. The aerofoil geometry and flap configuration were modelled using CATIA V5, and a Computational Fluid Dynamics (CFD) approach was employed for the analysis using ANSYS Fluent. Aerodynamic parameters, including lift coefficient, drag coefficient, and pressure distribution, were evaluated at various angles of attack to assess the performance improvements. The results of this study provide insights into the design optimization of aerofoils for high-lift applications in aircraft wings.

DOI: https://doi.org/10.5281/zenodo.19552524

Security Vulnerability Assessment And Risk Analysis_200

Authors: Akash Sharma, Arslaan, Shikha Sharma

Abstract: This study presents a systematic approach to vulnerability assessment and risk analysis within a controlled laboratory environment. A virtual network infrastructure was deployed, comprising Kali Linux as the scanning platform and Metasploitable 2 as the target system, to emulate a small-scale enterprise network. Network reconnaissance was conducted using Nmap, followed by vulnerability assessment using Nessus. Identified vulnerabilities were evaluated and classified based on severity using the Common Vulnerability Scoring System (CVSS), and subsequently mapped to corresponding risk levels. The analysis revealed multiple high-severity vulnerabilities, including the presence of default credentials and outdated services, which pose significant security risks and necessitate immediate remediation. Furthermore, the results underscore the effectiveness and extensive coverage of Nessus, supported by its comprehensive plugin database exceeding 80,000 entries. The proposed methodology provides a practical and reproducible framework applicable to both academic research and real-world cybersecurity assessments.

Determining The Statistical Measures On Student’s Performance Metrics Using Python Code

Authors: Dr Preetha V

Abstract: This research investigates the student’s performance metrics using statistical measures based on the dataset. Students studies and achievements have been distracted by various factors such as sleeping hours, internet access, motivational level and other environmental factors. The study of this research will be helpful to find the exact reasons for the performance of the students. Further, this research focusses on the python coding using pandas library to find the statistical measures. The research explored various datasets and find the best suitable dataset with 6000 entries for the exact analysis.

DOI: https://doi.org/10.5281/zenodo.19554687

Research On Classification Models In Machine Learning For IRIS Dataset

Authors: Dr Preetha V

Abstract: Clasification models in machine learning is used in various applications for prediction. It is a supervised learning model where the models can be trained with the required outputs. There are various types of classifiers such as Naïve Bayes, KNN, Decision trees and support vector machines. Neural network machine learning model is also used for classification and prediction. The IRIS dataset is most familiar and commonly used dataset for classification to predict the accuracy. The research focusses on the analysis of the KNN based classification and Naïve bayes classification algorithms for the performance analysis.

DOI: https://doi.org/10.5281/zenodo.19554708

UNI ASSIST-AI: An Intelligent University Assistant Chatbot with LLM, RAG, and Multimodal Capabilities

Authors: Riddham Kothari, Professor Anusha Marda

Abstract: University support systems are under increasing pressure to handle high volumes of student queries accurately and at scale. Traditional rule-based chatbots are rigid and brittle, while large language model (LLM)-based systems, though fluent, are prone to hallucination. This paper presents UNI ASSISTAI, a Retrieval- Augmented Generation (RAG)-based intelligent university assistant that grounds every generated response in verified institutional knowledge. The system integrates a semantic vector retrieval pipeline with a GPT-based generative model, and extends it with multimodal input capabilities—supporting text, voice (via ASR), and image (via OCR) queries. The backend is served through a FastAPI interface, and the frontend is implemented in React with TypeScript and Tailwind CSS. Experimental evaluation on a curated university FAQ and policy corpus yields a Precision of 0.87, Recall of 0.84, and F1-score of 0.85, outperforming both rule-based and vanilla LLM baselines. This work demonstrates that domain-specific RAG architectures offer a scalable, reliable path to academic AI assistants.

DOI: http://doi.org/

An AI-Driven Framework For Real-Time Disaster Detection Using Social Media Data With Geo-Spatial And Sentiment Analysis

Authors: Ms.Suma Chamundeswari, Siva Hareesh Raja, Telukuntla Manikanta Reddy, Neela Harika, Adhikari Swamy Naidu, Paramata Kasava Satya Sai Arjuna Reddy

Abstract: In recent years, social media platforms have become an important source of real-time information during natural disasters and emergency situations. Millions of users share posts, images, and location information that can provide valuable insights for disaster monitoring and response. However, identifying relevant disaster-related information from the massive volume of social media data remains a significant challenge. This paper presents an AI-driven disaster detection framework that utilizes social media analytics, location intelligence, and sentiment analysis to monitor and identify disaster events in real time. The proposed system collects social media posts and processes them using natural language processing and machine learning techniques to detect disaster-related content. Location intelligence methods are applied to extract geographical information from posts, enabling accurate identification of affected areas. In addition, sentiment analysis is used to evaluate public emotions and urgency levels associated with disaster events. The integrated framework helps emergency response teams gain situational awareness and make timely decisions during critical situations. Experimental evaluation demonstrates that the proposed approach effectively identifies disaster-related posts and provides meaningful insights for disaster management systems. The framework can support authorities and emergency organizations in improving response strategies and enhancing public safety.

DOI: http://doi.org/10.61463/ijset.vol.14.issue2.184

Intelligent Phishing Website Detection Using Machine Learning And URL Feature Analysis

Authors: Mrs. V. Suvarna, Mallidi Mohana Sudha, Gunturi Satyasai Phani Amrutha Sri Varshini, Mungara Shakthi Indra Varma, Gangapatrula Ram Karthikeyan, Nalli Prema Chandhu

Abstract: Phishing attacks have become one of the most common cybersecurity threats, targeting users by creating fraudulent websites that mimic legitimate platforms to steal sensitive information such as login credentials, financial data, and personal identity details. Traditional phishing detection approaches, such as blacklist-based systems and manual verification methods, are often inefficient and unable to detect newly emerging phishing websites in real time. Therefore, intelligent and automated detection mechanisms are required to improve cybersecurity and protect users from online fraud. This study proposes an efficient machine learning–based framework for detecting phishing websites using URL and domain-based features. The proposed system utilizes a dataset containing both legitimate and phishing website URLs collected from publicly available repositories. Data preprocessing techniques are applied to clean and normalize the dataset, ensuring consistency and improving model performance. Multiple machine learning algorithms including Logistic Regression, Decision Tree, Random Forest, AdaBoost, and Gradient Boosting are implemented and evaluated using stratified cross-validation techniques to ensure reliable prediction results. Among the evaluated models, ensemble learning algorithms demonstrate superior performance due to their ability to combine multiple weak learners and reduce prediction errors. In particular, the Random Forest classifier achieves high detection accuracy by analyzing key URL characteristics such as domain name structure, prefix and suffix usage, DNS records, URL length, and IP address patterns. The experimental results show that the ensemble model effectively distinguishes between legitimate and phishing websites with high accuracy, precision, recall, and F1-score.Furthermore, feature importance analysis is performed to identify the most influential attributes contributing to phishing detection, enabling better understanding of model behaviour and improving system transparency. The proposed framework provides a scalable and automated solution for detecting malicious websites, helping users identify fraudulent URLs before interacting with them. Overall, the proposed machine learning framework enhances phishing detection capability, improves cybersecurity awareness, and provides an efficient tool for protecting users against online phishing attacks.

DOI: http://doi.org/10.61463/ijset.vol.14.issue2.185

Brain Tumour Detection And Multiclass Classification Using Ensemble Deep Learning And Vision Transformers With Explainable AI

Authors: Mrs. D. Chakra Satya Tulasi, Neelam Sree Amrutha, Patamsetty C S R Srija, Pilli Karthik Kumar, Dondapati Rakesh, Balabhadruni L V H S Surya Gopal

Abstract: Early detection and accurate classification of brain tumours are critical for effective treatment planning and improved patient survival. Magnetic Resonance Imaging (MRI) is widely used for brain tumour diagnosis; however, manual inspection of MRI scans by medical experts is time-consuming and may produce inconsistent results due to variations in human interpretation. To address these limitations, this study proposes an automated deep learning framework for brain tumour detection and multiclass classification using MRI images. The proposed system leverages transfer learning with several pre-trained Convolutional Neural Network (CNN) architectures, including VGG16, VGG19, ResNet50, InceptionV3, InceptionResNetV2, and Xception, to extract meaningful features from MRI images. A dataset containing 3264 MRI images across four categories—Glioma tumour, Meningioma tumour, Pituitary tumour, and No tumour—is utilized, and data augmentation techniques are applied to increase the dataset size and improve model generalization. Based on experimental performance, the three best-performing models—VGG16, InceptionV3, and Xception—are integrated into an ensemble model named IVX16, which combines predictions to enhance classification accuracy and reduce overfitting. In addition, Vision Transformer (ViT) based models such as SWIN, Compact Convolutional Transformer (CCT), and External Attention Network (EANet) are implemented for comparative analysis. To improve transparency and reliability in medical decision-making, Explainable Artificial Intelligence (XAI) techniques, specifically Local Interpretable Model-Agnostic Explanations (LIME), are applied to highlight the tumour-affected regions in MRI images and validate model predictions. Experimental results demonstrate that the proposed ensemble framework achieves superior performance compared to individual deep learning models. Overall, the proposed approach provides an accurate, reliable, and explainable solution for automated brain tumour detection and classification, which can assist healthcare professionals in faster and more consistent clinical diagnosis.

DOI: http://doi.org/10.61463/ijset.vol.14.issue2.186

Deep Learning-Based Generation And Detection Of Face Morphing Attacks For Secure Biometric Authentication

Authors: Mrs. A. Daiva Krupa Nirmala, Karrothu Nikhitha, Mohammed Sultana, Pati Yuktha, Palika Pavan Sai, Uppalapati Rajesh

Abstract: The rapid adoption of biometric authentication systems, particularly facial recognition technologies, has significantly improved identity verification in applications such as border control, digital identity management, and secure access systems. However, these systems remain vulnerable to sophisticated biometric attacks, among which face morphing attacks pose a serious security threat. In a morphing attack, facial images of two or more individuals are digitally combined to create a synthetic image that resembles multiple identities, allowing attackers to bypass biometric verification systems. Detecting such manipulated images is challenging due to variations in illumination, facial expressions, accessories, and image quality. This study proposes a robust deep learning–based framework for the generation and detection of face morphing attacks in biometric systems. The proposed approach integrates an advanced feature extraction mechanism with machine learning–based classification techniques to effectively distinguish between genuine and morphed facial images. To enhance detection performance, image preprocessing and enhancement techniques are incorporated to reduce noise and improve feature representation. Additionally, a diverse morph dataset containing both Morph-2 and Morph-3 images is utilized to simulate realistic morphing attack scenarios and improve model generalization across different facial characteristics. Multiple experimental evaluations are conducted using several publicly available facial image databases. The performance of the proposed model is assessed using evaluation metrics such as accuracy, precision, recall, F1-score, and detection error rates. Experimental results demonstrate that the proposed framework significantly improves morphing attack detection accuracy and provides a reliable defence mechanism for biometric authentication systems. By enhancing detection reliability and robustness, the proposed approach contributes to strengthening the security of modern facial recognition systems against identity fraud and biometric spoofing attacks.

DOI: http://doi.org/10.61463/ijset.vol.14.issue2.187

A Transfer Learning-Based Mobile Net Framework For Automated Structural Crack Detection

Authors: Mr. B. Janu Naik, Balla Vudaya Naga Varshitha, Goluguri Venkata Jeethendra Reddy, Rayavarapu Soma Shekar, Rejeti Bharath Kumar, Voleti Surya Vasanth Krishna Prasad

Abstract: Cracks in infrastructure pose serious risks to public safety and require timely detection for effective maintenance. This study presents Deep Crack, a deep learning-based approach for image-based crack prediction. The proposed method utilizes Convolutional Neural Networks (CNNs) with the Rfcn_b architecture as the backbone, combined with transfer learning to improve detection performance. Extensive data preprocessing techniques, including image augmentation, are applied to address data limitations and enhance model generalization. The model is trained and validated on a diverse dataset, enabling it to accurately distinguish between cracked and non-cracked images. A customized classification layer, incorporating global average pooling and fully connected layers, is integrated to further improve performance. The effectiveness of the model is evaluated using metrics such as accuracy, precision, and recall, along with confusion matrix analysis and classification reports. Experimental results demonstrate that the proposed approach achieves high classification performance, making it suitable for real-world infrastructure monitoring applications. This work highlights the effectiveness of combining deep learning and transfer learning techniques for automated crack detection and emphasizes their potential applications in civil engineering and infrastructure management.

DOI: http://doi.org/10.61463/ijset.vol.14.issue2.188

EcomAI: Machine Learning–Driven Capital Estimation And Profitability Forecasting For Next-Generation E-Commerce Ventures

Authors: Mrs. K. Harika, Padala Sri Sai Harshitha, Cheekatla Sri Bindu Patvika, Sree Bala Damisetti, Palla Rambabu, Azhar Syed

Abstract: E-commerce start-ups are rapidly expanding in the digital economy, yet accurate estimation of initial investment and prediction of profitability remain critical challenges. This paper presents an enhanced data-driven framework that utilizes machine learning techniques to estimate start-up capital requirements and forecast future profitability. A structured dataset comprising key business indicators such as operational costs, marketing expenditure, infrastructure investment, and revenue-related factors is constructed and analysed. A regression-based predictive model is developed to identify relationships between these variables and financial outcomes. The proposed approach emphasizes effective data preprocessing, including normalization and outlier handling, to improve model reliability. Experimental evaluation demonstrates that the model is capable of extracting meaningful patterns and providing practical insights for financial planning. The results highlight the importance of feature influence in determining capital requirements and profit margins. This study contributes to the domain of intelligent business analytics by offering a scalable and interpretable solution that supports entrepreneurs and investors in making informed decisions. The integration of machine learning into financial forecasting enhances strategic planning and promotes sustainable growth in the competitive e-commerce landscape.

DOI: http://doi.org/10.61463/ijset.vol.14.issue2.189

Boosting Ensemble Machine Learning Approach For Porosity Prediction In Carbon Dioxide Storage Reservoirs

Authors: Mrs.K.Tulya Sree Simla, Manne Namratha Sai, Mantri Bala Subrahmanyam, Ompolu Janu Priyanka, Korubilli Manoj Kumar, Garaga Manikyam

Abstract: Accurate estimation of reservoir porosity is a critical factor in evaluating geological formations for carbon dioxide (CO₂) storage in carbon capture and storage (CCS) projects. Porosity directly influences the storage capacity and injectivity of subsurface reservoirs, making its accurate prediction essential for effective CO₂ sequestration planning. Traditional porosity estimation methods based on core analysis are reliable but often expensive, time-consuming, and limited in spatial coverage. With the increasing availability of well-log data, machine learning techniques provide an efficient data-driven alternative for predicting reservoir properties. This study proposes a machine learning–based framework for porosity prediction using boosting ensemble algorithms to support CO₂ storage assessment. Well-log data collected from the Mena Murtee-1 well in the Darling Basin, Australia, are used as input features, while laboratory-corrected porosity values serve as the target variable. Data preprocessing techniques are applied to remove noise, handle missing values, and eliminate multicollinearity among input parameters. Ensemble boosting algorithms including AdaBoost Regression, Gradient Boost Regression, and Extreme Gradient Boost Regression (XGBoost) are implemented and evaluated using standard statistical performance metrics. Experimental results demonstrate that boosting ensemble algorithms effectively capture complex non-linear relationships between well-log parameters and porosity values. Among the evaluated models, Extreme Gradient Boost Regression achieves the highest prediction accuracy and provides reliable porosity estimates for subsurface formations. The proposed framework enhances reservoir characterization accuracy and supports efficient evaluation of geological formations for carbon dioxide storage.

DOI: http://doi.org/10.61463/ijset.vol.14.issue2.190

Smart Railway Track Monitoring System Using Image Processing And Fuzzy Logic For Early Fault Detection

Authors: Mr. V. Hemanth Sai, Vaddi Harika, Ganapareddy Manasa Sree Vyshnavi, Avasarala Deepthi Vineetha, Kunche Vinay Kumar, Koyya Manohar Yadav

Abstract: Railway transportation plays a vital role in modern infrastructure by providing efficient and reliable movement of passengers and goods. However, railway track faults such as cracks, misalignments, and structural damage can lead to severe accidents, service disruptions, and significant economic losses if not detected at an early stage. Traditional railway track inspection methods mainly rely on manual monitoring and scheduled maintenance procedures, which are time-consuming, labour-intensive, and prone to human error. With the growing expansion of railway networks and increasing train speeds, there is a strong need for intelligent and automated systems capable of detecting track faults accurately and efficiently. This study proposes an automated railway track fault detection framework based on image processing and fuzzy logic techniques. The proposed system utilizes a vision-based approach in which images of railway tracks are captured using an embedded camera system and processed to identify potential defects. Image preprocessing techniques such as grayscale conversion, noise filtering, and segmentation are applied to enhance the quality of captured images and isolate important track features. Edge detection and thresholding methods are used to identify cracks or abnormalities present on the railway track surface. This intelligent classification approach helps reduce false detections while improving decision-making accuracy. Experimental results demonstrate that the proposed system can effectively identify track faults and provide reliable early warnings for railway maintenance teams. By combining image processing techniques with fuzzy logic-based decision support, the proposed framework enhances railway safety by enabling automated, real-time track inspection. The system can significantly reduce manual inspection effort, improve fault detection accuracy, and support proactive maintenance strategies for modern railway infrastructure.

DOI: http://doi.org/10.61463/ijset.vol.14.issue2.191

AI-Powered Early Brake Anomaly Detection With Explainable Predictive Intelligence

Authors: Mrs. N. Nikitha, Vepada Yagnesh, Jenna Meghanadh, Pothabattula Sowmya Sravanthi, V V Naga Raju Neetipalli, Koyyana Rajesh Vardhan

Abstract: This study proposes a secure and efficient machine learning-based framework for predicting brake failures in heavy commercial vehicles. In modern transportation systems, the Air Pressure System (APS) of heavy vehicles is continuously monitored using IoT-based sensors, which generate large volumes of operational data. Manually detecting brake faults from such large and highly imbalanced datasets is both time-consuming and inefficient. To address these challenges, the proposed approach utilizes K-Nearest Neighbour (KNN) imputation to handle missing data and Synthetic Minority Oversampling Technique (SMOTE) to manage class imbalance. Various machine learning algorithms, including Logistic Regression, Decision Tree, Support Vector Machine, Gradient Boosting, and Random Forest, are implemented and evaluated using stratified cross-validation techniques. Experimental results indicate that the Random Forest classifier achieves superior performance in terms of accuracy, precision, recall, F1-score, and ROC-AUC. To improve interpretability and build trust in the prediction process, Explainable Artificial Intelligence (XAI) techniques such as SHAP and LIME are incorporated, enabling clear understanding of model decisions. Additionally, feature selection methods are applied to reduce computational complexity while maintaining high prediction accuracy. The proposed framework enhances the reliability of brake fault detection, minimizes maintenance costs, and supports predictive maintenance strategies in heavy transport systems.

DOI: http://doi.org/10.61463/ijset.vol.14.issue2.192

Radio AI: A Machine Learning-Based Framework For Optimized Radiotherapy Treatment Planning

Authors: Ms. Y. Suma Chamundeswari, Angara Devi Mahi, Sanapala Pradeep, Bonthu Prasad, Talabhaktula Janaki Sriram, Mamidipalli Lovely Saimahesh

Abstract: Radiotherapy treatment planning is a vital component in modern cancer management, requiring precise delivery of radiation to tumour regions while preserving surrounding healthy tissues. Conventional planning approaches are often manual, time-intensive, and limited in their ability to adapt to patient-specific variations. To address these challenges, this study explores a machine learning-driven framework for intelligent radiotherapy planning. The proposed approach leverages advanced deep learning architectures, particularly Convolutional Neural Networks (CNNs), along with classical machine learning models such as Support Vector Machines (SVM) and Random Forests (RF), to enhance tumour segmentation, dose estimation, and treatment optimization. By utilizing multimodal medical imaging data, including CT, MRI, and PET scans, the system enables accurate identification of tumour boundaries and supports data-driven clinical decisions. Furthermore, the integration of techniques such as multimodal learning and reinforcement-based optimization improves the adaptability and precision of treatment planning. The results demonstrate that the proposed framework achieves high segmentation accuracy and reliable dose prediction, contributing to improved treatment effectiveness and reduced adverse effects. This work highlights the transformative potential of machine learning in enabling personalized, efficient, and intelligent radiotherapy solutions.

DOI: http://doi.org/10.61463/ijset.vol.14.issue2.193

Enhancing Autonomous Vehicle Security Using Explainable Artificial Intelligence For Anomaly Detection

Authors: Mrs. Ch. Veera Gayathri, Bhaviri Sri Ganesha Seetha Hanuma Gowri, Sangani Praveen Dhana Kumar, Ryali Yuvaraj, Ganta Venkata Sridhar, Gadhi Subrahmanya Krishna Teja

Abstract: Autonomous driving systems have emerged as a transformative technology in modern intelligent transportation, enabling vehicles to operate with minimal or no human intervention. These systems rely heavily on large volumes of sensor data, communication networks, and machine learning algorithms to make real-time driving decisions. However, the increasing integration of autonomous vehicles into vehicular networks has also introduced significant cybersecurity and safety challenges. In particular, anomalous behaviours caused by cyber-attacks, faulty sensors, or malicious vehicles in Vehicular Ad Hoc Networks (VANETs) can threaten the reliability and safety of autonomous driving environments. Detecting such anomalies using traditional monitoring approaches is difficult due to the complexity, scale, and dynamic nature of vehicular communication data. To address these challenges, this study proposes an explainable artificial intelligence (XAI)–based anomaly detection framework for autonomous driving systems. The proposed framework integrates machine learning models with explainability techniques to identify abnormal behaviours in vehicular networks while also providing transparent interpretations of model decisions. Initially, autonomous driving datasets are pre-processed through feature extraction, redundancy elimination, data balancing, and normalization to improve model performance. Several machine learning algorithms, including Decision Tree, Random Forest, Support Vector Machine (SVM), K-Nearest Neighbour (KNN), Deep Neural Network (DNN), and AdaBoost, are implemented to classify vehicles as normal or anomalous based on their behavioural features. To enhance interpretability, the framework incorporates explainable AI techniques such as SHapley Additive explanations (SHAP) and Local Interpretable Model-Agnostic Explanations (LIME). These methods provide both global and local explanations by identifying the most influential features contributing to anomaly detection decisions.

DOI: http://doi.org/10.61463/ijset.vol.14.issue2.194

Automated Rice Grain Quality Assessment Using Computer Vision And Machine Learning Based On Physical Feature Extraction

Authors: Mr.K.Srikanth, Nulu Subrahmanya Venkata Lakshmi, Kadmati Amrutha Varshini, Devara Sneha, Gangireddla Rahul Sai Manikanta, Karri Durga Prasad

Abstract: Rice quality assessment plays a critical role in the agricultural and food industries, as it directly influences market value, consumer satisfaction, and food safety. Traditionally, rice grain quality evaluation has been performed manually by experienced inspectors based on visual observation and physical measurements. However, manual inspection is often time-consuming, subjective, and prone to human error. To overcome these limitations, this study proposes an automated rice quality analysis system using image processing and machine learning techniques. The proposed approach extracts important physical attributes of rice grains, including area, perimeter, width, height, aspect ratio, and major and minor axes, from digital images captured under controlled conditions. Image processing techniques such as grayscale conversion, binary thresholding, morphological operations, edge detection, and object detection are applied to accurately isolate and measure individual rice grains. The extracted features are then stored and used to train a machine learning model for classification. A Support Vector Machine (SVM) classifier is employed to categorize rice grains into different quality grades based on their physical characteristics. The performance of the proposed system is evaluated using a dataset consisting of multiple rice varieties. Experimental results demonstrate that the automated system achieves improved classification accuracy compared to traditional manual inspection methods. The proposed framework provides an efficient, reliable, and cost-effective solution for automated rice quality assessment. By integrating computer vision and machine learning techniques, the system reduces human dependency, improves consistency in quality grading, and has the potential to support large-scale deployment in agricultural industries and food processing units.

DOI: http://doi.org/10.61463/ijset.vol.14.issue2.195

CapsuleVision: An Interpretable Deep Learning Framework For Wireless Capsule Endoscopy Image Classification

Authors: Mrs. N. V. S.Sowjanya, Dammala Bhanu Durgesh, Chodavarapu Sriram, Vara Bhanu Prasad, Penta Rameswar, Sheik Shameerulla

Abstract: Deep learning has significantly advanced medical imaging and computer-aided diagnosis (CAD), enabling accurate disease detection. However, the limited interpretability of deep learning (DL) models restricts their clinical adoption. To address this, Explainable Artificial Intelligence (XAI) techniques are used to better understand model decisions. In endoscopic imaging, diagnosis is mainly based on manual visual inspection, which can be time-consuming and subjective. Integrating automated DL systems can improve both accuracy and efficiency. In this study, multiple transfer learning models are applied to a balanced subset of the Kvasir-Capsule dataset, consisting of the top nine classes. The Vision Transformer (ViT) achieves the best performance with an F1-score of 97% ± 1%, outperforming existing approaches. Other models, including MobileNetV3Large and ResNet152V2, also achieve F1-scores above 90%.To enhance interpretability, XAI techniques such as Grad-CAM, Grad-CAM++, Layer-CAM, LIME, and SHAP are used to generate heatmaps highlighting important regions in the images. These visual explanations provide insights into model decisions and reduce the black-box nature of DL models. Overall, this work combines high accuracy with improved transparency, contributing to more reliable and trustworthy medical AI systems.

DOI: http://doi.org/10.61463/ijset.vol.14.issue2.196

Intelligent Fire Detection And Early Warning System Using Deep Learning And Computer Vision

Authors: Mr. M. V. Rajesh, Adapa Amrutha Varshini, Sabbarapu Kumar Ganesh, Mutyala Sowmya, Koppisetti N V Prasanna Sri Sandeep, Kanupudi Mahathi Sree

Abstract: Early fire detection plays a vital role in preventing major disasters, reducing property loss, and ensuring public safety. Conventional fire detection systems primarily depend on physical sensors such as smoke, heat, and gas detectors. While these methods are commonly used, they often face challenges such as delayed response, high false alarm rates, and reduced effectiveness in complex environments like industrial areas and densely populated urban regions. With the rapid growth of computer vision and deep learning technologies, image-based intelligent fire detection systems have emerged as a promising solution for early detection. This study presents a deep learning-based fire detection and early warning system that utilizes Convolutional Neural Networks (CNN) to automatically detect fire from images captured through surveillance cameras. The proposed model extracts visual features from images and classifies them into two categories: fire and non-fire. A well-structured dataset consisting of fire and non-fire images is used for training and validation of the model. To enhance generalization and minimize overfitting, data augmentation techniques such as rotation, scaling, and horizontal flipping are applied. Additionally, optimization methods including Early Stopping and ReduceLROnPlateau are incorporated to improve training efficiency and model stability. The experimental findings indicate that the CNN-based model performs significantly better than traditional machine learning approaches such as Logistic Regression, K-Nearest Neighbor (KNN), and AdaBoost. The proposed system achieves high classification accuracy along with strong recall and AUC values. Moreover, an automated alert mechanism is integrated into the system, which triggers an alarm upon detecting fire, enabling quick emergency response. Overall, the proposed approach offers a cost-effective, reliable, and scalable solution for fire detection, suitable for deployment in surveillance systems across buildings, industrial sectors, and smart city environments. The results demonstrate that deep learning-based visual fire detection systems can greatly improve safety monitoring and disaster prevention.

DOI: http://doi.org/10.61463/ijset.vol.14.issue2.197

Review On Smart Digital IC Testing System Using Microcontroller

Authors: Atharv Santosh Shinde, Mahesh Mahadev Patil, Omkar Shashikant Salunkhe, Ajay Sampat Khade, Onkar Mohan Maske, Ganesh Yashwant Nangare, Mrs. Sima Shinde

Abstract: Integrated Circuits (ICs) are the foundation of modern electronics; however, manual functional verification is often labour-intensive and prone to errors. This paper introduces a Smart Digital IC Testing System that utilizes an Arduino Mega 2560 microcontroller to automate the testing of 74-series logic gates (AND, OR, NOT, NAND, NOR, XOR), a 555 timer, and a 741 op-amp through a ZIF socket, a 4×4 keypad, and an LCD/buzzer for feedback. The system applies predefined input patterns and compares the outputs against embedded reference values, providing pass/fail results. It incorporates protective features such as current-limiting resistors and voltage clamping. Priced at ₹6040, this system offers high accuracy, portability, and upgradability, making it ideal for laboratories and repair shops. It effectively addresses the shortcomings of previous Arduino/ by integrating mixed-signal verification within an affordable platform.

DOI:

E-Commerce Product Management And Shopping Portal

Authors: D. Sowjan, Ch. Keerthi Sri, B. V V S S Lakshmi Prasanna, M. V Vara Prasad, Mrs. Ch. Vijaya Lakshmi

Abstract: In this paper, we designed and developed an e-commerce system a future-friendly E-Commerce Product Management and Shopping Portal utilizing the latest technologies, including the MERN stack (MongoDB, Express, React, and Node.js), Vision AI, and Blockchain to make the system work more efficiently work and user experience are presented. The suggested system will receive a dual-interface construction that will serve both administrators and end users with specific features. In our case, the admin portal was developed to efficiently manage product management operations. It is possible to easily add new items, modify products and organize them in one place. Moreover, Vision AI was applied to automate the process of analyzing photos in order to extract information regarding the properties of products including color, type, and category. At the same time, inventory management is implemented, which allows for real-time updates on the stock of products and alerts concerning the shortage of items. Apart from that, admins have an opportunity to control customers' orders, change the status of these orders and analyze sales reports to make decisions based on them. As for the end-users of the website, they can easily browse products, look for them, put them into the cart and buy goods using the provided checkout page. To increase trust, Blockchain was applied for the purposes of ensuring users about the security of transactions. Specifically, once users complete their orders, their data will be securely stored and not be available for any further modifications by other users. The described technologies help create an efficient and convenient system.

Artificial Intelligence In Early Disease Detection: Trends, Applications, And Challenges

Authors: Dadi Jonathan Abba, Mafeng Jamima Dudari, Raliyah Umar Alkaleri, Habibu Aminu Sani, Kudyo Deborah Yona

Abstract: By improving diagnostic precision, lowering clinician workload, and facilitating early illness identification, artificial intelligence (AI) is transforming healthcare. In order to enhance patient outcomes, cut mortality, and save healthcare expenses, early illness detection is essential. AI processes massive medical datasets, evaluates medical pictures, and helps physicians make clinical decisions by utilizing machine learning, deep learning, and predictive analytics. With a focus on applications in cancer, cardiology, neurology, infectious illnesses, and personalized medicine, this study explores current developments in AI-assisted diagnostics. The advantages AI offers medical practitioners are also covered, such as increased patient monitoring, less mistakes, and higher diagnosis accuracy. Despite these benefits, issues including algorithmic bias, high implementation costs, data privacy, inadequate physician training, and ethical considerations continue to be major obstacles. In order to maximize AI adoption in early illness diagnosis and healthcare delivery, the study concludes by examining future views and highlighting the necessity of cooperative research, policy frameworks, and integration techniques.

DOI: https://doi.org/10.5281/zenodo.19565314

Smartfarm Price Advisor: A MACHINE LEARNING Framework For Agricultural Price Estimation

Authors: Sivaranjani S, Kanishka B, Mr Sankaran D

Abstract: Thus, agricultural price volatility is found to affect farmers’ income, agricultural markets, and policy planning. Price forecasting helps agricultural producers make informed decisions. In this context, the paper proposes a SmartFarm Price Advisor framework that utilizes machine learning techniques to accurately forecast commodity prices based on past modal, minimum, maximum prices, and rate of change. This paper analyzes commodity prices for a multi-year time frame, preprocesses the data to obtain extensive insights, and produces exploratory results like commodity price trends, statistical summaries of individual commodities, and correlation structures between variables. Various supervised learning and time series techniques are implemented and studied to find the most appropriate methodology suitable for forecasting commodity prices for different crop categories. From the experimental implementation, the correlations between key commodity prices are found to be strong, while significant commodity-specific changes are observed. Therefore, the proposed commodity price system has great potential to be implemented in agricultural settings.

Multimodal Deep Learning For Respiratory Disease Prediction Using Lung Sounds And Chest Images

Authors: Ashmi Jomon

Abstract: Pneumonia remains one of the leading causes of mortality worldwide, necessitating accurate and timely diagnostic tools. Conventional diagnostic approaches often rely on a single modality such as chest X-rays or CT scans, each providing valuable but distinct clinical information. This paper presents a multimodal deep learning framework that integrates three com-plementary diagnostic modalities—Chest X-Ray images, Chest CT Scan images, and Lung Sound audio recordings—for robust and flexible pneumonia detection. Three independent deep learning models are developed: a DenseNet121 architecture for Chest X-Ray classification, a ResNet50 architecture for CT Scan analysis, and a custom Convolutional Neural Network (CNN) for Lung Sound classi-fication, where raw audio recordings are converted into Mel spectrogram images prior to inference. An attention-based Late Fusion mechanism dynamically combines the probability outputs of the individual models by assigning learned trust weights to each modality through an attention network and producing a final consensus prediction via a dedicated consensus network. The complete system is deployed as a Flask-based web applica-tion supporting both single-modality and comprehensive multi-modal prediction modes, enabling adaptability across different clinical scenarios. Experimental evaluation demonstrates that the proposed system effectively supports reliable predictions across individual modalities while also enabling enhanced inference when multiple modalities are available, evaluated using standard metrics including Accuracy, Precision, Recall, F1-Score, and ROC-AUC. The system demonstrates significant potential as an accessible and clinically meaningful decision support tool for early pneu-monia detection.

OneSec: Real-Time Email Phishing & Threat Detection System

Authors: Prabhakaran S, Nishal R, Thillaiarasu J, Dr. M. Rajesh Babu

Abstract: Phishing attacks remain one of the most economically devastating cyber threats, accounting for approximately 91% of all cyberattacks according to the Anti-Phishing Working Group (APWG). Despite advances in enterprise-grade solutions, accessible, interpretable, and privacy-preserving tools for individual Gmail users are critically lacking. This paper presents OneSec, a full- stack web application that integrates with the Gmail API via OAuth 2.0 read-only access to proactively detect phishing emails in real time before the user opens them. The system employs a seven-rule, weighted multi-factor threat engine that evaluates IP-based URL usage, suspicious top-level domains (TLDs), SPF/DKIM authentication failures, reply-to header anomalies, credential-harvesting keywords, social-engineering urgency patterns, and excessive URL density. Real-time threat alerts are delivered via Server-Sent Events (SSE) to a React 18 TypeScript dashboard with sub-500 ms end-to-end latency. Empirical evaluation on a balanced 500-email benchmark (250 PhishTank phishing, 250 Enron/legitimate) yields precision of 91.25%, recall of 87.60%, F1-score of 89.4%, and a mean detection latency of 340 ms. User acceptance testing achieves a System Usability Scale (SUS) score of 82.5, rated Excellent. OneSec is open-source, self-hostable at zero cost, and requires no machine learning infrastructure, making advanced phishing protection accessible to all Gmail users.

DOI:

AQUA FUEL

Authors: Divy Rohela, Devansh Rohela, Adarsh Tiwari, Megha Thakur, Harshita Verma, Aman Ali, Prof.Ashok Soni

Abstract: This research explores the feasibility of operating a small-scale internal combustion engine using hydrogen-rich gas produced through an onboard water electrolysis system. The generated gas, commonly referred to as HHO (a stoichiometric mixture of hydrogen and oxygen), is obtained by electrochemically splitting water using an alkaline electrolyte-based setup powered by an external electrical source. The produced hydrogen–oxygen mixture is directly supplied to a modified 100 cc four-stroke spark ignition engine, replacing conventional hydrocarbon-based fuel. To enable this operation, significant modifications were made to the carburetor and intake manifold to ensure controlled gaseous fuel delivery, proper mixing with intake air, and stable combustion conditions. Hydrogen’s unique combustion characteristics—such as very low ignition energy, high flame propagation speed, and wide flammability range—require careful control of air–fuel ratios and ignition timing. Experimental results indicate that the engine can operate successfully under hydrogen-rich conditions, demonstrating the technical feasibility of onboard hydrogen generation for small engine applications. The exhaust emissions show a major reduction in carbon-based pollutants, with water vapor being the dominant byproduct. However, the system’s overall performance is strongly influenced by electrolysis efficiency, electrical energy consumption, gas production rate, and thermal stability of the engine. Key limitations identified include high power demand for electrolysis, risk of backfire, heat accumulation in intake components, and the need for precise flow control mechanisms. Despite these challenges, the study demonstrates that onboard hydrogen generation can serve as a transitional clean-energy solution for internal combustion engines in small-scale applications.

DOI:

Cryptocurrency as an Investment Asset: A Study of Bitcoin and Ethereum

Authors: Assistant Professor Dr. Shweta Oza

Abstract: The growth of the crypto markets has changed the investment environment in a profound manner by elevating cryptocurrencies from purely speculative assets to institutional-grade investments. The current paper evaluates the investment characteristics of Bitcoin and Ethereum, the most popular cryptocurrencies, based on the modern portfolio theory framework. According to the analysis carried out for 2020-2025, the Bitcoin asset demonstrates an impressive Sharpe ratio of 1.7, substantially exceeding that of the S&P 500 (0.54) and gold (0.48-0.54). In favorable market conditions, Ethereum outperforms Bitcoin in terms of risk-adjusted returns, exhibiting even better characteristics. The study highlights a change in the mechanism of price fluctuations in the market from the "four-year cycle" to the flow of institutional capital. At the same time, correlation analysis shows that despite the absence of high correlation of these assets with other asset classes over the long term, the correlation between the two increases under market pressure. As a result, 1-4% of portfolio weight can be safely allocated to each asset, depending on the investment strategy.

DOI: http://doi.org/10.5281/zenodo.375

Digital Marketing Trends And Consumer Buying Behavior: An Empirical Study

Authors: Dr. Viji R, Director, Prof. (Dr.) Vellayan Srinivasan, Dr.V.O.Kavitha

Abstract: Abstract- Digital marketing has transformed the way businesses interact with consumers, significantly influencing buying behavior across industries. This study examines the impact of current digital marketing trends—such as social media marketing, influencer marketing, personalized advertising, and search engine optimization—on consumer buying behavior. A quantitative research approach was adopted with a sample of 120 respondents from urban and semi-urban areas. The findings reveal that social media marketing and influencer endorsements have the strongest influence on purchasing decisions, while email marketing and banner ads show moderate impact. The study concludes that digital marketing strategies significantly shape consumer perception, trust, and purchase intention.

DOI: https://doi.org/10.5281/zenodo.19591009

Design And Implementation Of A Real-Time Event And Emergency Management System(BANDOBAST)

Authors: Utkarsh Suryawanshi, Pratik Ghodke, Siddheshwar Mankar, Vitthal Kamble

Abstract: This paper presents a scalable and secure real- time event and emergency management system developed using Flask, Flutter, and Firebase. Traditional systems rely on manual coordination, resulting in inefficiencies and delayedresponse times. The proposed system integrates mobile and web technologies to enable real-time tracking, task management, and emergency response. The system also incorporates clustering techniques to optimize resource allocation. Experimental results demonstrate improved coordination, faster response time, and enhanced operational efficiency.

Decentralized Storage System: A Novel Approach To Secure And Resilient Data Management

Authors: Aaditya Chomal, Snehil Gamit, Sidharth Panda, Aqsa Rangrez, Govind, Yassir Farooqui

Abstract: In the modern digital landscape, traditional cen- tralized storage models are increasingly vulnerable to security breaches, suffer from single points of failure, incur high main- tenance costs, and present scalability limitations. The Decen- tralized Storage System (DSS) is proposed as an alternative solution, utilizing distributed ledger technologies, peer-to-peer (P2P) networks, and advanced cryptographic mechanisms to establish a fault-tolerant, secure, and highly available data storage infrastructure. This paper presents the design and imple- mentation of a decentralized storage framework that integrates key blockchain concepts—such as immutability, transparency, and consensus validation—to enhance data integrity and security. The system enables users to store, retrieve, and share data in a fully decentralized manner while ensuring confidentiality through encryption. The work includes a comprehensive architectural and functional analysis of a full-stack, decentralized file storage platform built specifically on the Filecoin Protocol, leveraging the InterPlanetary File System (IPFS) for distributed content addressing and efficient data retrieval. The platform employs a hybrid architecture combining Web2 technologies—Next.js for the frontend, Node.js/Express for the backend, and MongoDB for centralized metadata management—with core Web3 protocols. The analysis confirms the project’s success in creating a practical, non-custodial storage solution that abstracts the complexities of the decentralized web. However, a key architectural trade-off is identified: the system’s reliance on provider-centric tooling (Boost CLI) and third-party Remote Procedure Call (RPC) endpoints (Glif API) simplifies development but introduces dependencies that compromise the ideal of full, end-to-end decentralization.

Design And Implementation Of A Low-Cost IoT-Based Vehicle Tracking System Using ESP32 And GPS Technology

Authors: Abdulmalik O. Usman, Ibrahim A. Barde

Abstract: With increasing demand for real-time vehicle tracking, cloud-based solutions are being developed, integrating IoT technologies with web applications. This study proposes a complete vehicle tracking setup using Firebase for real-time data synchronisation and OpenStreetMap for the visualisation of vehicle location. With an ESP32 microcontroller, the vehicle forwards GPS data, both latitude and longitude, to a cloud database; users then can monitor the movements of the vehicle through a dynamic web interface. Important features include secure user authentication, geofencing capabilities, and instant notifications on predefined events. Solutions to problems regarding data accessibility and system scalability provide evidence that IoT-enabled vehicle tracking systems can improve public safety and assist law enforcement. The results suggest that the integration of these hardware and software technologies would be instrumental in formulating solutions for challenges in modern-day security.

DOI: https://doi.org/10.5281/zenodo.19592420

Attention Augmented Multi-Branch Architecture For Early Eye Disease Diagnosis

Authors: T Sreenivasu (Sr.Asst. Professor), Pranitha J, P Saptagiri, S shara, Ch Sudha, M Akshitha

Abstract: The human eye is a critical sensory organ, and any impairment in its function can significantly affect an individual’s quality of life. Eye diseases such as glaucoma, cataract, and retinal disorders can lead to severe vision loss if not detected at an early stage, making timely and accurate diagnosis essential for effective treatment and prevention. In this study, an automated eye disease detection system was developed to address the limitations of existing approaches, including insufficient feature representation, lack of interpretability, and high computational complexity. A comprehensive preprocessing strategy was employed to enhance image quality and improve robustness against variations in input data. The proposed approach effectively learned robust and discriminative features for accurate classification of multiple eye diseases while maintaining computational efficiency. In addition, a visualization technique was incorporated to highlight the important regions influencing the model’s predictions, thereby improving transparency and supporting better clinical interpretation, which can contribute to enhanced diagnostic confidence and overall clinical performance. The system was trained and evaluated on a multi-class eye disease dataset and demonstrated consistent improvement in classification performance and interpretability compared to conventional methods. The integration of efficient feature learning and visual interpretability enhances the reliability and practical applicability of the system, making it a promising solution for real-world computer-aided eye disease diagnosis.

DOI: https://doi.org/10.5281/zenodo.19594539

Application Of Constructed Wetlands With Modified Substrate In Produced Water Treatment

Authors: Erewari Ukoha-Onuoha, Iyechikame Silvernus

Abstract: Produced water also known as oilfield brine is one of the largest waste by volume produced during oil and gas exploration activities. Its high volume, complex composition, environmental and health hazards makes it one of the most critical waste in the waste stream of the oil and gas industry. Most wastewater treatment methods employed in the treatment of PW are often energy-intensive, costly, and generate harmful by-products. With more stringent regulations on produced water handling and disposal, there is the need for the application of more efficient yet cost effective, less energy intensive and minimal by-products systems. One of such system is the constructed wetland that mimics nature. To this end, this study investigated the performance of a constructed wetland system with modified substrates for the treatment of produced water, focusing on the removal of heavy metals (Cd, Pb, and Zn) and oil and grease (O&G). A batch scale, three (3) constructed wetland experimental setup arranged in series was established. Gravel, coarse sand, loamy soil, and biochar were used as substrates. Phragmites australis, a common reed species was planted in each constructed wetland. Heavy metals were analyzed using Atomic Absorption Spectrophotometer while oil and grease was analyzed by gravimetric method. Results of the study demonstrated appreciable removal efficiencies for heavy metals (Pb: 82%, Cd: 75%, Zn: 74%) and O&G (86%) within a short hydraulic retention time of 45 minutes. This suggest that constructed wetland with a biochar layer can serve as an effective and eco-friendly treatment method for produced water.

DOI: https://doi.org/10.5281/zenodo.19602319

The Influence Of Landscape Layout And Lighting On Perceived Safety In Higher Institutions; Caleb University As A Case Study

Authors: Ademakinwa Olasunmbo, Fagbola Oluwateniolafunmi D, Oke Athalia A, Owolabi Emmanuel K

Abstract: Campus safety at Caleb University is influenced by environmental design factors such as landscape layout and lighting. This study examines their impact on students’ safety perception using a mixed-methods approach, combining site observations with questionnaire data from 300 students. Findings show that while daytime safety is generally high, nighttime safety is lower due to poor lighting, dark spots, and obstructed visibility from vegetation. The study concludes that improving lighting, maintaining landscapes, and enhancing spatial planning are essential for safer campus environments, providing practical recommendations for design interventions in institutional settings.

DOI: https://doi.org/10.5281/zenodo.19602878

SmartFinops: Industry 4.0 Expense Intelligence Powered By AI For Industrial Financial Tracking

Authors: Dr. K. Ravikiran, Margam Manisha, Spoortika Boora, Advala Kalyani

Abstract: Abstract. SmartFinOps is an automated financial process tool that can be used to achieve the highest efficiency and accuracy through the use of artificial intelligence and based on the core ideas of Smart Manufacturing and Industry 4.0. The SmartFinOps platform will integrate AI functions with TensorFlow.js to automatically extract expense data, intelligentcategories, anomalydetection and prediction of financial analysis. The system utilizes cloud-based infrastructure to use it together with real-time analytics to ensure that the manual financial tracking process is altered to a smart and flexible process in a digital workflow. The platform is developed through React 19 and Next.js 15, includes safe authentication supported by Clerk and NextAuth, has a reliable scaleand dependability, and manages access securely. Supabase is used to store real-time data in an organized manner using Prisma ORM,and interactive dashboards can display the visual representation of the spending behaviours and financial trends. SmartFinOps other characteristics are voice-based logging, automatic PDF Reports, and AI-assisted suggestions based on this technology are useful in enhancing interaction between machines and humans. SmartFinOps enables organizations to become faster in the speed at which they respond and enhance their operational performance through the application of the Smart Manufacturing principles to develop Intelligent Automation, Convert Information to Real-time Monitoring, and Bring Digital Transformation to Financial Operations Management.

Secure And Transparent E-Voting System Using Blockchain Technology

Authors: I. Janaki Devi, M. Prathap, M. Ravi Teja, T. Durga Prasad, Mr. S.N V P Ravi Teja

Abstract: The use of block chain technology in the Online Voting System with Face Recognition would go a long way in improving the security, transparency, and integrity of the election process. Through block chain, all the votes made by a registered voter can be stored as a transaction in an unchangeable and immutable ledger that can guarantee the safety and auditability of the data. The voter registration information, encrypted Voter IDs, Aadhaar numbers, and hashed face recognition data could be safely placed on privately-owned block chain instead of on centralized data bases that can be at risk of attack. Such a decentralized system is free of the risks of data manipulation and guarantees the credibility of voter data during the election process. The tokenization aspect of Block chain means that a voter is allocated unique voting tokens so that they cannot cast multiple votes. The tokens are verified and stored in the block chain, which increases the election security even more. Smart contracts may be used to automate a number of processes in the election management process, including counting votes and announcing the winner, which provides transparency and prevents manipulation. The final election outcome may also be recorded in the block chain that end up being available to the populace and anyone able to confirm the integrity of the election process. Based on the use of block chain and React JS in frontend, and Spring Boot in backend to deal with API requests and MySQL to store non-sensitive information, the system establishes a secure, decentralized, and fully auditable election system that facilitates transparency, privacy, and fraud deterrence.

Smart Hire AI Based Multi-Round Hiring Proctoring System

Authors: J. Durga Dinesh, T.Kalyhann, P.Samitha, G.Teja Sri, Parasa Sivadurgarao

Abstract: The project develops a cloud-based interview management system which uses artificial intelligence to improve recruitment processes for both candidates and recruiters. The system provides automated hiring processes which begin from application submission and continue through evaluation while using webcam interviews and AI-powered question creation. The portal allows candidates to register and upload their resumes while they can also control their profiles and take MCQ and coding and HR interviews which the system secures through locked tabs during testing to stop cheating. The recruiters can examine resumes and coordinate interview times while they create intelligent questions through the Gemini AI API and distribute invitation links through email. The system provides automatic response assessment together with instant selection outcome delivery. The platform uses Spring boot with java for backend processing ReactJS for frontend development and MySQL for secure data storage to maintain efficient and accurate and fair hiring processes.

A Comprehensive Study On Quality Assurance Strategies And Advanced Testing Techniques In Modern Web Development

Authors: Suresh Karena, Jay Patel, Priyanshu Mishra, Parth Acharya, Om Dave, vaibhavi parikh

Abstract: With the rapid evolution of web technologies, ensuring the quality, reliability, and security of web applications has become increasingly complex. Quality Assurance (QA) plays a critical role in maintaining software standards by integrating systematic testing practices throughout the Software Development Life Cycle (SDLC). This paper presents a comprehensive analysis of QA methodologies, testing techniques, and automation frameworks in modern web development.

Enhancing Multilingual Machine Translation Using Context Aware Large Language Models

Authors: Ritik Sadh, Preeti Sharma, Priyanshu Singh, Vansh Guleria, Akthar Warsi

Abstract: Multilingual is a critical component of global communication systems. Despite significant (NMT), contextual ambiguity, low-resource language, domain adaptation persist. Enhanced by leveraging context-aware (LLMs). By integrating transformer-based architectures with contextual embeddings, the proposed approach improves semantic consistency, translation fluency, and cross-lingual transfer learning. The study BLEU and METEOR while also considering qualitative human evaluation. Results indicate that context-aware LLMs significantly outperform traditional models in handling long-range dependencies and multilingual tasks. The paper concludes with a discussion on limitations and future research directions.

A Comparative Study of Traditional vs. Digital Advertising Effectiveness

Authors: Shiva Shri G, Associate Professor Dr. T. M. Hemalatha

Abstract: Advertising plays a crucial role in influencing consumer awareness and purchase decisions. With rapid technological advancements, businesses now use both traditional advertising methods such as television, radio, newspapers, and billboards, as well as digital advertising platforms like social media, websites, and online streaming services. This study aims to compare the effectiveness of traditional and digital advertising in influencing consumer behavior. The research adopts a descriptive research design using both primary and secondary data. Primary data is collected through a structured questionnaire from 100 respondents, while secondary data is sourced from books, journals, and previous studies. Statistical tools such as percentage analysis, Chi-square test, and ANOVA are used for data analysis. The study helps businesses understand consumer preferences and supports the need for an integrated advertising approach. Furthermore, the study highlights the growing importance of digital platforms in reaching a wider and more targeted audience, while also recognizing the continued relevance of traditional advertising in building trust and brand credibility. By analyzing consumer responses and preferences, the research provides insights that can help businesses choose the most effective advertising strategy and optimize their marketing efforts in a competitive environment.

DOI: https://zenodo.org/records/19606234

E-Commerce Growth and Challenges: A Study of Trends, Issues and Prospects

Authors: Rishi Durairaj.P, Associate Professor Dr. T. M. Hemalatha

Abstract: businesses work and how people buy things. The quick spread of digital technology, more people using the internet, and the rise of smartphones have all helped online shopping grow a lot. This study looks at how e-commerce is growing, along with the main trends, difficulties, and possible future developments in the industry. The research focuses on consumer behavior, technological developments, payment systems, and logistical aspects influencing the growth of online retail platforms. The study adopts a descriptive research design and relies on both primary and secondary data sources. The findings reveal that while e-commerce provides convenience, wider product availability, and competitive pricing, it also faces several challenges such as cybersecurity risks, logistical constraints, and issues related to consumer trust. The study concludes that addressing these challenges and strengthening digital infrastructure will enhance the sustainability and future growth of the e-commerce sector.

DOI: https://zenodo.org/records/19606621

A Study of Fintech and Digital Banking: Reshaping Traditional Banking

Authors: Ms:B.Kareshmithra, Associate Professor Dr. T. M. Hemalatha

Abstract: This study examines the impact of Financial Technology (FinTech) and digital banking on traditional banking systems. With the rapid growth of services such as UPI, mobile wallets, internet banking, and online transactions, banking operations have become more technology- driven and customer-centric. The research aims to understand customer adoption patterns, benefits, and challenges associated with digital banking. A descriptive research design is adopted, and data will be collected through structured questionnaires from bank customers. Secondary data is gathered from journals, reports, and online sources. The study seeks to analyze how FinTech is transforming traditional banking operations and influencing customer behavior. The findings will provide insights for banks to enhance digital services while maintaining security and customer trust.

DOI: https://zenodo.org/records/19606851

Electric Vehicle (EV) Charging Parameters Estimation On A Web Portal

Authors: Manjeet Kumar Maurya, Prashant Shukla, Pal Devendra, Poorvi Srivastava

Abstract: This paper proposes a comprehensive EV charging portal built on the MERN stack, addressing infrastructure gaps, range anxiety, and operational inefficiencies in emerging markets like India. The system integrates real-time station discovery, secure bookings, payments, and ML-driven estimation of critical EV parameters—State of Charge (SOC), State of Health (SOH), charging efficiency, and range prediction. Our React frontend delivers interactive geospatial maps and dashboards, while Node.js/Express backend handles REST/WebSocket APIs connected to OCPP-enabled chargers. Deployed on AWS Kubernetes, it scales to 1K+ concurrent users with <200ms latency. Results show 95% SOC estimation accuracy (MAE: 2.1%) and 98% booking success rate, validated against real-world datasets. This work provides a practical blueprint for scalable EV ecosystems.

DOI: https://doi.org/10.5281/zenodo.19606871

The Role of Digital Marketing in Business Growth

Authors: Vel Murugan P, Assistant Professor Ms. Bushra B

Abstract: Digital marketing has emerged as a crucial driver of business growth in the modern economy. With the increasing penetration of the internet and digital technologies, businesses are shifting from traditional marketing methods to digital platforms to reach a wider audience. This study examines the role of digital marketing in enhancing business growth by analyzing its impact on customer engagement, brand awareness, and revenue generation. The research adopts a descriptive methodology using primary and secondary data. The findings reveal that digital marketing significantly contributes to business expansion by improving customer reach, enabling targeted marketing, and increasing conversion rates. The study emphasizes the need for businesses to adopt innovative digital strategies to sustain competitive advantage.

DOI: https://zenodo.org/records/19607068

Digital Payment Systems and Their Impact on Consumer Behavior

Authors: Athithya.R, Assistant Professor Ms. Bushra B

Abstract: Digital payment systems have become one of the most influential innovations in modern financial transactions. With the rapid development of mobile banking, Unified Payments Interface (UPI), digital wallets, debit/credit cards, QR-code payments, and contactless payment methods, consumers have gradually shifted away from cash-based transactions. The growth of digital payment platforms has significantly influenced consumer behavior by changing purchasing habits, payment preferences, spending frequency, trust levels, and convenience expectations. This study examines the impact of digital payment systems on consumer behavior by analyzing the adoption patterns, factors influencing consumer preference, and the behavioral changes created by digital transaction services. The research focuses on key areas such as ease of use, security perception, trust in online transactions, frequency of digital payments, and consumer satisfaction. A descriptive research design was adopted, and data was collected using a structured questionnaire supported by secondary sources. Percentage analysis, Chi-square test, and One-way ANOVA were applied to interpret consumer behavior patterns. The findings show that digital payment systems have improved transaction speed, reduced dependency on physical cash, increased online purchase activity, and enhanced consumer satisfaction. However, challenges such as fraud risks, technical failures, and lack of awareness among certain groups remain barriers. The study concludes that digital payment systems play a major role in shaping consumer behavior and recommends improved security frameworks, awareness campaigns, and better infrastructure support for sustained adoption.

DOI: https://zenodo.org/records/19607333

AI-Driven Intelligent Shopping Recommendation And Analytics Platform Using Natural Language Processing

Authors: Purvi Pal, Rishabh Raj, Krrish Nayak, Mrs. Geetha C

Abstract: Contemporary e-commerce ecosystems confront a structural personalization deficit: conventional keyword-centric retrieval systems cannot interpret intent-rich natural language queries, while static rule-governed recommendation engines fail to capture the dynamic behavioral signals that reveal genuine user preferences. This paper presents the design, implementation, and evaluation of an AI-Driven Intelligent Shopping Recommendation and Analytics Platform that resolves both deficiencies by integrating a transformer-grounded semantic search engine with a behaviorally adaptive collaborative filtering recommendation module. The semantic search component transforms free-text user queries into 768-dimensional dense vector embeddings via a Gemini-powered embedding pipeline and executes approximate- nearest-neighbor retrieval against a MongoDB Atlas Vector Search index, augmented by LLM-extracted structured filter predicates covering price ceiling, minimum star rating, and color attribute. The recommendation engine assigns differential weights to heterogeneous interaction signals—product views, selection clicks, cart inclusions, and completed transactions—aggregating them into continuously updated per-user preference vectors. A scalable three-tier deployment architecture cleanly partitions the React.js presentation layer, the Flask API processing layer, and the MongoDB persistence layer, enabling independent horizontal scaling of AI inference components. Comprehensive testing confirms semantic search response times within two seconds and recommendation feed generation within three seconds, substantially outperforming traditional keyword-based baselines while operating entirely on open-source infrastructure at a fraction of the cost of commercial personalization services.

AI Driven Customer Churn Prediction For Telecom Indusrty Using Machine Learning Algorithms

Authors: P.Sri Sai Srujan, I.Babeeswara Reddy, K.Vinay Raj, Mrs.Geetha C

Abstract: Customer churn is a major challenge for telecom companies as losing customers directly affects revenue. Artificial Intelligence (AI) and Machine Learning (ML) techniques can analyze telecom customer data and predict which customers are likely to discontinue services. This study proposes an AI-driven churn prediction system using machine learning algorithms such as Logistic Regression, Decision Tree, Random Forest, and Gradient Boosting. The model analyzes customer usage behavior, billing information, and service interactions to identify patterns associated with churn. The results show that machine learning models can effectively detect high-risk customers, enabling telecom companies to implement proactive retention strategies and improve customer satisfaction.

A Performance Assessment Of Machine Learning-Based Techniques For Image Restoration

Authors: Shradha Kumavat, Kapil Shah

Abstract: Image restoration is a fundamental task in image processing with wide-ranging applications in modern life, including medical imaging, remote sensing, radar imaging, and digital preservation of historical and museum artifacts. The objective of image restoration is to recover a high-quality image from degraded observation by reducing the effects of noise and blur. Effective restoration depends on understanding the degradation process; therefore, identifying the type of noise and the blur model is essential. In practical scenarios, images are often degraded by atmospheric and environmental conditions and restoring them requires appropriate restoration techniques tailored to the distortion characteristics. This paper reviews and assesses contemporary machine learning-based image restoration methods. The proposed evaluation reports quantitative performance across four standard benchmark datasets Kodak24, CBSD68, Urban100, and LIVE—using PSNR (dB), MSE, and SSIM as primary quality metrics. The achieved PSNR scores are 27.24 dB, 29.38 dB, 30.04 dB, and 30.91 dB on Kodak24, CBSD68, Urban100, and LIVE, respectively. The corresponding MSE values are 367.56, 224.88, 193.10, and 158.02, while SSIM values are 0.8690, 0.9337, 0.9432, and 0.8008. These results demonstrate the effectiveness of the evaluated approach in improving image quality across diverse image restoration benchmarks.

DOI:

A Systematic Review Routing Protocols In Wireless Sensor Networks Priya Yadav

Authors: Priya Yadav

Abstract: Wireless Sensor Networks (WSNs) have become an essential technology for monitoring and data collection in various domains such as environmental monitoring, healthcare, military surveillance, and smart cities. Routing protocols play a crucial role in WSNs because sensor nodes have limited energy, processing power, and communication capabilities. Efficient routing mechanisms are required to ensure reliable data transmission while minimizing energy consumption and prolonging network lifetime. This paper presents a systematic review of routing protocols in wireless sensor networks. The study categorizes routing protocols into different types such as data-centric, hierarchical, and location-based routing protocols. Key protocols including LEACH, PEGASIS, TEEN, and Directed Diffusion are analyzed in terms of energy efficiency, scalability, and performance. The paper also discusses challenges and future research directions for improving routing efficiency in WSNs.

Rent Management System

Authors: Vaishnav Devanshu

Abstract: The traditional rent collection and management system at Prayashmay Hospitality Pvt. Ltd. was inefficient, error-prone, and time-consuming. This paper presents an automated Rent Management System that streamlines tracking, rent collection, and financial reporting using digital payments and a centralized dashboard. The system eliminates manual errors, enhances transparency, and provides real-time financial insights. The research discusses the challenges of manual rent collection, the methodology used for system implementation, and the impact on financial efficiency and decision-making.

Performance Analysis Of An Intrusion Detection System Based On Big Data Analytics And Ensemble Techniques.

Authors: Ayodeji Ireti Fasiku, Oghenerukevwe Oyinloye

Abstract: Datasets encompass a wide range of network activities and intrusion patterns. The traditional intrusion detection systems (IDS) are struggling to provide all-round protection to the network but unable to analyze the new volumes of data and the velocity of today’s networks. This research leverages the capabilities of big data analytics to process and analyze large-scale datasets collected from network traffic logs. Feature engineering and selection techniques were applied to extract relevant features that capture the distinguishing characteristics of normal and intrusive activities. Each model in the ensemble is trained independently using a subset of the data, utilizing their unique algorithms and strengths. The proposed system employs a range of machine learning models including Support Vector Machines (SVM), Decision Trees, Naive Bayes, k-Nearest Neighbors (KNN), Random Forest, Neural Networks, and two ensemble techniques, Bagging Ensemble, and XGBoosting. A comprehensive comparative analysis of these models were conducted to evaluate their performance in detecting intrusions accurately and efficiently. Hence, a comparative analysis was carried out to evaluate the performance of each model individually and as part of the ensemble. Performance metrics such as accuracy, precision, recall, F1-score, and area under the receiver operating characteristic curve (AUC-ROC) are employed to assess the effectiveness of the models in identifying intrusions and minimizing false positives. The research contributes to the field of intrusion detection by providing insights into the performance of different machine learning models when applied to big data analytics and ensemble techniques. The comparative analysis aids in selecting the most effective models for building robust IDS solutions, improving network security, and safeguarding critical information assets against emerging cyber threats.

DOI: https://doi.org/10.5281/zenodo.19610733

Optimization Of Machining Parameter For Aluminium Material

Authors: Mr. Satyendra Kundaram, Mr. Akshay Khose, Mr. Dnyaneshwar Kale, Mr. Sai Vidhate, Dr. Yadav M.S, prof. Dilip Kupade, Mr. Maner Anis M

Abstract: Machining is an important manufacturing process used to shape metal materials into desired sizes and shapes. Aluminium is widely used in industries because of its light weight, good strength, and corrosion resistance. However, proper selection of machining parameters is necessary to achieve good surface finish, longer tool life, and higher productivity. This project focuses on the optimization of machining parameters such as cutting speed, feed rate, and depth of cut during the machining of aluminium material. Different machining experiments are conducted by changing these parameters to study their effect on surface roughness and material removal rate. The Taguchi method is used as a statistical technique to analyze the results and determine the best combination of parameters. The main objective of this project is to improve machining performance, reduce production cost, and increase efficiency in aluminium machining operations. The optimized parameters obtained from this study can help industries achieve better quality products with minimum machining time and tool wear.

DOI: https://doi.org/10.5281/zenodo.19614884

CRDT Based Peer To Peer Collaboration Editor

Authors: Abhinav Anil, Rajan Patel, Nilay Pravin Sabnis, Mrs Geetha C

Abstract: The design of this product presents a fully featured collaborative document editor built using Conflict-free Repli-cated Data Types (CRDTs) and peer-to-peer communication over WebRTC. The system leverages Y.js as its core synchro-nization engine to enable real-time, decentralized collaboration without reliance on a centralized server for document state management. Traditional collaborative editors often depend on Operational Transformation (OT) and centralized architectures to maintain consistency. In contrast, this implementation adopts a CRDT-based model, allowing each client to maintain a local replica of the document while guaranteeing eventual consistency across all peers. By integrating WebRTC for direct peer-to-peer communication, the system reduces latency, improves scalability, and enhances resilience against single points of failure. The editor supports rich-text formatting, concurrent multi-user editing, cursor awareness, offline editing with au-tomatic synchronization upon reconnection, and conflict-free merging of changes. Y.js manages shared data structures and efficiently propagates incremental updates between peers. The WebRTC layer establishes secure data channels for real-time message exchange, enabling seamless synchronization across distributed clients. Key challenges addressed include peer dis-covery, network reliability, conflict resolution in highly concur-rent environments, and maintaining low-latency performance at scale. The architecture separates document state management, networking, and user interface layers to ensure modularity and extensibility. This method demonstrates how CRDTs combined with modern web technologies can deliver a scalable, decentral-ized alternative to traditional collaborative editing systems. The resulting platform provides strong consistency guarantees, high performance, and fault tolerance, making it suitable for real-time document collaboration in distributed and intermittently connected environments.

DOI: https://doi.org/10.5281/zenodo.19615256

AI- Based Smart Home Automation System

Authors: Prof. Mr.Sunny Wasudeorao Thakare, Rohit Kumar

Abstract: This project presents a smart home automation system that combines Artificial Intelligence (AI) with Internet of Things (IoT) technologies to improve comfort, safety, and energy efficiency. The system uses an Arduino-based controller connected with multiple sensors to monitor environmental conditions such as temperature, smoke, and water levels. Based on the collected data, the AI logic processes inputs and automatically controls home appliances like fans, lights, doors, and pumps. For example, cooling devices are adjusted according to temperature, and alerts are generated during abnormal situations like fire or gas leakage. Additionally, a face recognition feature is used to enhance home security by allowing access only to authorized individuals. This system demonstrates how intelligent automation can simplify daily activities and create a safer living environment.

Development Of An Adaptive E-Learning System Using LLM Modules

Authors: Abhishek Shrivastava, Philip Christopher, Nikhil Singh

Abstract: The rapid development of artificial intelligence (AI) provides new opportunities for improving e-learning environments. AI-powered Moodle modules can transform traditional learning management systems into adaptive, personalized platforms that deliver digital content tailored to the needs of learners. These modules focus on automated test generation, intelligent learning, predictive analytics, and learning content creation, which can reduce educator workload and improve student engagement. Our pilot experiments show that integrating AI into Moodle can increase efficiency, learning outcomes, and overall satisfaction for both learners and educators.

Study On Fire Alarm System

Authors: Sujal Gongale, Amardip Raipure, Saurabh Zade, Ujwal Ratnaparkhi, Prof. Deepali Vaidya mam.

Abstract: A fire alarm system using a flame sensor is an advanced safety system designed to detect fire at an early stage by sensing infrared (IR) radiation emitted by flames. Unlike traditional smoke or heat detectors, flame sensors provide faster detection, especially in open environments. This paper discusses the working principle, components, advantages, and applications of flame sensor-based fire alarm systems.A fire alarm system using a flame sensor is an advanced safety system designed to detect fire at an early stage by sensing infrared (IR) radiation emitted by flames. Unlike traditional smoke or heat detectors, flame sensors provide faster detection, especially in open environments. This paper discusses the working principle, components, advantages, and applications of flame sensor-based fire alarm systems.

DOI: https://doi.org/10.5281/zenodo.19626707

AI Integrated Landmine Detection System

Authors: Sahil Ravi Telote, Om Sanjay Haspatil, Shreyash Sharad More, Prof. Sanjay N. Jadhav

Abstract: The AI-Based Travel Itinerary System is an intel- ligent web application designed to simplify trip planning by automatically generating personalized travel itineraries based on user preferences. The system uses Artificial Intelligence to analyze inputs such as destination, travel duration, budget, and interests. It integrates APIs like Google Maps and AI models to provide real-time recommendations including routes, attractions, and accommodations. The system reduces manual effort and enhances travel planning efficiency by generating optimized schedules and cost estimations.

DOI: https://doi.org/10.5281/zenodo.19627677

Securing Mobile Banking Ecosystems: An Integrated Survey Of Cryptographic Standards, Authentication Protocols, And Fraud Mitigation

Authors: Anvar Sadath A K, Hafeesa M Habeeb

Abstract: Mobile banking has emerged as the dominant channel for financial services, but its rapid adoption has also made it a prime target for cyberattacks. Despite the availability of strong cryptographic primitives and advanced authentication models, real-world implementations remain plagued by misconfigurations, API vulnerabilities, and human factors. This survey provides a structured technical review across three foundational pillars of secure mobile banking: encryption standards, authentication protocols, and risk mitigation models. We analyze the evolution from legacy schemes such as RSA and SMS OTPs to modern approaches including TLS 1.3, ECC, FIDO2, and AI-driven fraud detection. Comparative analysis reveals that while cryptographic algorithms are robust in theory, weak deployments and usability–security trade-offs continue to undermine resilience. Real-world case studies—including SIM-swap fraud, banking malware (Zeus, Anubis, Cerberus), and OAuth misconfigurations—are used to contextualize threats. Finally, we synthesize research gaps such as lightweight quantum-resistant cryptography, explainable AI in fraud detection, and standardized API security, outlining a roadmap toward globally harmonized, user-centric, and adaptive mobile banking security frameworks.

Role Of AI In Modern Marketing

Authors: Ms: A. Divya, Ms. Bushra B

Abstract: Artificial Intelligence (AI) has emerged as a transformative force in modern marketing, reshaping how businesses analyze data, engage customers, and design strategic campaigns. Unlike traditional marketing approaches that relied heavily on generalized messaging and manual market research, AI-driven systems enable real-time data processing, predictive analytics, and hyper-personalization.By leveraging machine learning algorithms, natural language processing, and data mining techniques, organizations can understand consumer behavior patterns, forecast demand, and deliver customized experiences at scale. From recommendation engines used by digital platforms to automated customer service chatbots, AI enhances operational efficiency while improving customer satisfaction and brand loyalty.

DOI: https://doi.org/10.5281/zenodo.19628763

Emerging Trends In Smart Proctoring: A Comprehensive Review Of Machine Learning-Based Exam Supervision Systems

Authors: Shruthi S V, Chethan H K

Abstract: The imperative for robust academic integrity in the era of remote assessment has led to the development of Intelligent Exam Supervision (IES), commonly known as smart proctoring. This monograph provides an exhaustive analysis of the machine learning (ML) architectures and socio-technical frameworks necessary for building scalable, effective, and ethically compliant IES systems. Part I establishes the theoretical context, distinguishing between traditional and automated supervision, and examining the economic drivers for ML adoption. Part II delves into the core technological engine: the multimodal data pipeline. We detail the collection, synchronization, and fusion of heterogeneous streams—including high-resolution video biometrics, acoustic forensics, and low-latency keystroke dynamics—using advanced techniques like Temporal Convolutional Networks (TCNs) and Cross-Attention Transformers, exploring the challenges of real-time edge processing and sensor reliability. Part IV addresses the most critical domain: ethics, legal compliance, and fairness. This section extensively analyzes global regulatory frameworks (GDPR, BIPA, CCPA, and emerging frameworks in Asia-Pacific), the application of Adversarial Debiasing for algorithmic fairness, and the critical role of Explainable AI (XAI) in generating justifiable, transparent audit trails (SHAP, LIME), including the formal definition of the Cost of Misclassification and its policy implications. Part V explores the challenge of Adversarial Machine Learning (AML) and the use of Generative Adversarial Networks (GANs) for defense hardening and robust synthetic data generation. Part VII conducts a deep analysis of the Psychological and Pedagogical Impact on students, including the surveillance effect, the necessary curricular reform, and the detailed architecture of the Human-in-the-Loop (HITL) system. Finally, the work concludes by advocating for a holistic socio-technical design where technological innovation is inextricably linked to ethical governance and pedagogical necessity, alongside the security imperatives of Post-Quantum Cryptography.

DOI: https://doi.org/10.5281/zenodo.19629693

Paarsh Matrimony System

Authors: Tushar Kathe, Sapna Bhimajiyani

Abstract: Finding a suitable life partner has always been an important but challenging task. Traditional methods of matchmaking are often time-consuming, less accurate, and lack privacy. This research paper presents the Paarsh Matrimony System, a modern web-based application designed to improve the matchmaking process using technology and intelligent features. The system is built using React.js, FastAPI, and MongoDB, ensuring a smooth and responsive user experience. It includes features such as secure user authentication, profile management, AI-based recommendations, real-time chat, and payment integration. The platform focuses on providing accurate match suggestions based on user preferences and behavior. It also ensures data security and user privacy. The results show that the system makes matchmaking faster, easier, and more reliable compared to traditional methods.

Cryptographically Verifiable Retrieval-Augmented Generation: A Tripartite Architecture For Decentralized Provenance, Compute-to-Data Privacy, And Automated Fact-Checking

Authors: Sheetal Laroiya, Purnendu kumar Ghosh, Ketan, Subhanshu Raj

Abstract: Contemporary Large Language Models (LLMs) deployed within Retrieval-Augmented Generation (RAG) pipelines suffer from three distinct vulnerabilities: epistemological opacity (hallucinations), non-consensual data exploitation, and a lack of granular provenance. This paper proposes a tripartite architecture to resolve these deficits. First, we introduce an Indexer and Provenance Layer utilizing Decentralized Identifiers (DIDs), Verifiable Credentials (VCs), and on-chain mapping to establish immutable audit trails for retrieved context. Second, we present a Privacy-Preserving Compute-to-Data paradigm leveraging tokenized access control to facilitate economic incentivization without exposing raw data to consumers. Finally, we formalize a Verifiable RAG Pipeline equipped with a multi-strategy Verifier Agent to autonomously audit LLM generation against cryptographically anchored evidence.

A Robust Ensemble Machine Learning Framework For Accurate Sentiment Classification Of Twitter Data

Authors: Ritu Suryavanshi, Sharad Morolia

Abstract: Social media platforms generate a massive amount of opinion-based data that reflects public attitudes toward various topics such as politics, products, and social events. Among these platforms, Twitter is widely used for expressing opinions in the form of short textual messages known as tweets. Analyzing these tweets can provide valuable insights into public sentiment. However, sentiment classification of Twitter data is challenging due to informal language, abbreviations, emojis, and sarcasm. This study proposes an ensemble learning framework to improve the accuracy of Twitter sentiment classification. The framework involves several stages, including data collection, preprocessing, feature extraction using techniques such as Bag-of-Words and TF-IDF, and training multiple machine learning classifiers. Ensemble methods combine the predictions of these classifiers to generate more reliable results. The performance of the proposed model is evaluated using metrics such as accuracy, precision, recall, and F1-score. The proposed approach aims to enhance sentiment analysis performance and provide more accurate insights from social media data.

Missing Person Tracker Using AI And Face Recognition.

Authors: Payal N. Rathi, Anuj K. Bendre, Tanuja B. Bhand, Shivam D. Khodake

Abstract: Missing person cases are a serious global issue that require fast and accurate identification systems. Traditional methods rely on manual investigation, which is time-consuming and often inefficient. This paper presents an AI-based Missing Person Tracking System that uses face recognition techniques based on computer vision and deep learning [1].The system uses Convolutional Neural Networks (CNNs) to extract facial features and generate embeddings for accurate identification under different conditions such as lighting and facial expressions. OpenCV is used for face detection and preprocessing, [2] and the extracted features are converted into numerical vectors. These vectors are compared with a centralized database using similarity measures like Euclidean distance to find matches.The system also provides a user-friendly interface that allows users and authorities to upload images and retrieve results in real time. This approach improves automation, reduces manual effort, and increases search efficiency. Overall, the system offers a reliable and scalable solution for identifying missing persons and supporting public safety.

Blockchain For Cybersecurity: Safeguarding Data In The Digital Age

Authors: Sachin Kumar, Nishanth Bhaskar, Adinath M. Nair, Shaik Mohammad Gouse, Dr. Priya R. Swaminarayan

Abstract: The remarkably rapid and rapid growth of the internet and the incessant progress of digital technologies has utterly transformed nearly all aspects of life in modernity. It has however come along with significant vulnerabilities that it has been able to endanger safety and confidentiality for individual and organization and government information. Cyber-attacks, which may involve stealing money, could range from data theft and ransomware attacks to phishing, all forms of hacking are a threat this among other forms has become so common and more rampant. Such damages for the most part, extensive and vital for people, various enterprises, and communities. This paper discusses several aspects related to the integration of blockchain technology into the field of cybersecurity, particularly it focuses on its promising ability to enhance data integrity while protecting sensitive information that requires careful handling, and simultaneously provide strong and effective defenses in opposition to the continuously changing and evolving landscape of various cyber threats. This will go all the way using blockchain in applications such as identity management, IoT security, ransomware defense and integrity of data. An academic paper on the importance of how this the decentralized, cryptographically secured blockchain technology can transform cybersecurity frameworks. Further, this research attempts to address the limitation and challenges that exist including scalability issues, energy usage concerns that may accrue as well as those regulatory hurdles that may pose a hindrance to it. There has been wide acceptance of blockchain in the area of cybersecurity this would be possible through comprehensive review of literature, case Through comprehensive research and many practical implementations that have been done, this paper It provides a balanced view of the revolutionary. The potential of blockchain but full appreciation of the pragmatic the different forms its barriers to adoption take. The article ends with a lengthy, in-depth discussion of future prospects of the blockchain technology in Such involvement included what was considered as the need for continued cybersecurity-continued Research and technological advancements in the above problem. Present Limitations and Optimize Blockchain Technology This would bring effectiveness in the protection of digital assets in the digital age into the fore.

DOI:

Renal Ai: A Hybrid Convergent Ai Framework for Advanced Chronic Kidney Disease Detection and Diagnostic Support

Authors: Aiswarya Priyadharshini J M, Daksha R, M Anand

Abstract: RenalAI is an advanced AI-enhanced diagnostic ecosystem designed for the early and accurate detection of Chronic Kidney Disease (CKD). Manual diagnosis of renal pathologies is often delayed due to the separate analysis of clinical reports and medical imaging. To address this, the project proposes a Hybrid Convergent AI Framework that integrates two powerful models: a Convolutional Neural Network (CNN) for detecting abnormalities in CT scans and a Random Forest Classifier for risk stratification based on clinical biomarkers. The framework employs advanced preprocessing techniques to ensure data integrity across multi-modal inputs. By converging these two modules, RenalAI achieves a high diagnostic accuracy of 95.8%, providing a reliable second opinion for clinicians. Furthermore, the system incorporates a voice-enabled assistant using Speech Recognition and NLP to allow hands-free clinical navigation and automated report generation. RenalAI serves as a critical Diagnostic Support System, reducing human error and significantly improving patient outcomes through precision diagnostics and intelligent automation.

DOI: https://doi.org/10.5281/zenodo.19631677

Sterility Studies and Isolation of Bacteria from Vaccine Carriers used in Primary Health Care Centres in Awka, Anambra State

Authors: Okoli, F. A, Anieto, E.C, Anazodo C.A, Awari V.G., Okoye K.C, Ebo, P.U., Ogbunude, A.P, Chidozie, C.P, Okoye C. V., Orji, C.C, Obi Z.C.

Abstract: This study investigated the sterility of vaccine carriers used in selected Primary Health Care (PHC) centres in Awka, Anambra State, with the aim of assessing possible bacterial contamination and determining the antimicrobial susceptibility patterns of isolated organisms. Vaccine carriers are critical components of the cold chain system, ensuring that vaccines remain potent and effective during transportation and storage. However, poor handling, inadequate cleaning, and improper maintenance can compromise their sterility, thereby posing risks to public health. Samples collected from vaccine carriers were subjected to microbiological analysis. The isolates obtained included Staphylococcus aureus, Staphylococcus epidermidis, and Pseudomonas aeruginosa. Antimicrobial susceptibility testing revealed varying sensitivity patterns: S. aureus showed susceptibility to ciprofloxacin (15.2 mm), streptomycin (14.3 mm), levofloxacin (12.7 mm), rifampicin (12.2 mm), and gentamicin (11.8 mm). S. epidermidis was sensitive to chloramphenicol (15.8 mm), ciprofloxacin (14.3 mm), gentamicin (12.8 mm), rifampicin (11.5 mm), and ampiclox (10.1 mm). P. aeruginosa demonstrated susceptibility to augmentin (14.5 mm), ofloxacin (13.5 mm), and cefuroxime (12.7 mm). The presence of these organisms indicates that vaccine carriers, if not adequately sterilized, can serve as reservoirs for pathogenic bacteria capable of causing nosocomial and community-acquired infections. The results highlight the urgent need for strict adherence to cleaning protocols, regular training of health workers, and periodic sterility checks on vaccine carriers. Strengthening these practices will help safeguard vaccine integrity, maintain public confidence in immunization programs, and enhance overall disease prevention efforts

DOI: https://doi.org/10.5281/zenodo.19634211

Antifungal Potentials Of Ginger (Zingiber Officinale) On Human Pathogen Candida Species

Authors: Akpadolu C.B, Uwanta, L.I., Victor-Aduloju, A.T, Anazodo, C.A, Awari, V.G., Anieto, E.C, Ebo, P.U., Okoye, K.C, Ogbunude, A.P, Agu, K.C

Abstract: The increasing prevalence of antifungal-resistant Candida species, such as Candida albicans, Candida glabrata, and Candida parapsilosis, poses a significant challenge in managing oral and vaginal fungal infections in humans. This study investigates the antifungal potential of ginger (Zingiber officinale), a commonly used medicinal plant known for its antimicrobial properties, against these pathogenic Candida species in tropical Africa. Methanolic and aqueous extracts of ginger was examined for their antifungal activity, through preliminary antimicrobial screening, minimum inhibitory concentration (MIC), and minimum fungicidal concentration (MFC) assays. The species of candida isolated was described as F1 which aligns with Candida albicans, F2 with Candida parapsilosis, F3 with Candida krusei, and F4 with Candida galbrata. Candida albicans is known to produce germ tubes, while other Candida spp typically does not exhibit such structural characteristics. The methanolic extract was particularly effective against Candida albicans and Candida parapsilosis, while the aqueous extract was more potent against Candida krusei at a concentration of 75mg/ml and 100mg/ml with an inhibition zone diameter of 9.50mm and 18.00mm. The methanolic extract generally exhibited stronger antifungal activity than the aqueous extract suggesting its possible therapeutic application in treating candida associated infections.

DOI: https://doi.org/10.5281/zenodo.19634596

Anti-Theft Flooring Mat

Authors: Samruddhi Patil, Samruddhi Yawale, Manjusha Tatiya, Vidya Dhoke

Abstract: Besides all that, there is the IoT-Based Anti-Theft Mat System. As it is evident from the name, the security system is intended for securing the property from any unauthorized access by means of a relatively inexpensive apparatus, which employs a principle of the pressure detection method. Different from the current devices which utilize motion sensors and often suffer from false alarm, this system will detect a person's footprint. This concept is based on the employment of the pressure sensor installed inside the mat, data about which will be forwarded to the microcontroller. Once someone steps on the mat, the device will immediately notify its user about it through Wi-Fi connection. Thus, users will be notified about intrusion instantly once it happens. Due to the IoT, users will be able to receive notifications independently from their location at that moment. There are several benefits offered by the proposed system such as its cost-effectiveness, easy operation and availability. It will be easy for users to install such a device either in their office or residence since no expenses will be required for installation and purchase of additional hardware.

DOI: https://doi.org/10.5281/zenodo.19634796

CT Image-Based 3D Modeling And PatientSpecific Cranial Implant Design

Authors: Rubina Begam M, Nitishkumar G, Nandhagopalan E, Sanjiv S K, Tamilselvan D

Abstract: Cranial defects resulting from trauma, tumor resection, congenital disorders, or decompressive craniectomy present significant functional and aesthetic challenges. Traditional implant design relies on manual intraoperative shaping or prefabricated generic implants, leading to poor fitting, prolonged surgical time, and suboptimal outcomes. This study proposes a comprehensive digital workflow for patient-specific cranial implant design using Computed Tomography (CT) imaging and open-source 3D modeling tools. The methodology employs CT DICOM images processed through 3D Slicer for threshold-based segmentation (200-3000 HU), volumetric reconstruction, and surface mesh generation. Missing cranial regions are reconstructed using bilateral symmetry-based mirroring, and Boolean operations extract the precise implant geometry. The implant is refined in Meshmixer for thickness uniformity (2.5mm), edge beveling, and surface quality optimization before STL export. Experimental validation on five test cases with defect areas ranging from 15-85 cm² demonstrated mean alignment errors of 0.74mm (maximum 1.12mm), consistent implant thickness, and 100% watertight mesh topology. The complete workflow achieves processing times of 2.4-4.8 hours per case, representing a 70% reduction compared to commercial services, with zero software licensing costs. Results confirm that open-source CT-based workflows can achieve clinically acceptable geometric accuracy for patient-specific cranial reconstruction, making advanced implant design accessible to resource-limited healthcare settings.

Smart Patch for Continuous Biometric Data Collection and Analysis for Athlets

Authors: Dr.D. Karthikeyan, K. Anitha, K. Bhuvaneshwari, M. Charumathi

Abstract: Athlete identification using biometric data is an advanced approach that utilizes unique physiological and behavioural characteristics of individuals. Parameters such as heart rate, ECG signals, body temperature and motion patterns are continuously monitored to create a distinct identity profile for each athlete. Unlike traditional identification methods, biometric-based systems provide higher accuracy and security. Wearable devices and smart patches enable real-time data collection during training and competitions. The collected data is processed using microcontrollers and analysed using intelligent algorithms. These systems can recognize athletes based on their unique patterns and detect any abnormalities in performance or health. Wireless communication technologies allow seamless data transfer to mobile applications and cloud platforms. Coaches and trainers can monitor athlete performance and identity remotely through dashboards. This method also helps in preventing impersonation and ensuring fair participation in sports events. Overall, biometric-based athlete identification enhances both security and performance monitoring in modern sports environments. In addition, biometric-based identification systems provide continuous authentication, ensuring that the athlete is verified throughout the activity rather than at a single point in time. This improves security and reduces the chances of identity fraud in competitive sports. The system also supports personalized training by analysing individual performance trends over time. Integration.

IoT Based Non-Invasive Uric Acid Monitoring System Using Color Sensing And Cloud Analytics

Authors: Mrs. R. Aarthi, T.K. Harish Ragavendar, K. Annamalai, R. Adhithiyan, K. Abishek

Abstract: Uric acid is a key biomarker for diagnosing metabolic disorders such as gout and chronic kidney disease. Conventional lboratory methods are expensive, time-consuming, and require trained professionals, limiting accessibility in point-of-care settings. This paper presents a low-cost, portable, IoT-enabled uric acid detection system based on colorimetric analysis using the phosphotungstic acid (Folin's) method. The system employs a TCS34725 RGB color sensor interfaced with an ESP32 microcontroller via I2C protocol to measure the blue channel intensity of the Tungsten Blue compound formed during the reaction. The principle follows Beer-Lambert law, where absorbance is directly proportional to uric acid concentration. A 3D printed light-isolated chamber ensures consistent optical measurements by eliminating external light interference. The ESP32 processes sensor data using a calibration equation derived from standard solutions and transmits results wirelessly to Arduino IoT Cloud for real-time remote monitoring via web dashboard and mobile application. Validation was performed using bovine urine samples compared against certified laboratory analysis. The proposed system offers a simple, reliable, and affordable solution for point-of-care uric acid monitoring.

Codewiz – The Intelligent Coding Assis- Tant

Authors: Rathod Yash, Patel Riya, Mr.Biju Balakrishnan

Abstract: CodeWiz presents the design and implementation of a next-generation intelligent online coding platform that fuses contemporary web development technologies with advanced arti- ficial intelligence to deliver comprehensive, real-time support for coding and developer productivity. Unlike conventional code edi- tors that offer limited static functionalities, CodeWiz utilizes con- textual understanding and AI-driven analytics to convert raw code inputs and user interactions into actionable guidance, dy- namic suggestions, and personalized programming feedback. The system architecture adopts a modern MERN stack, using Mon- goDB for scalable data storage, Express.js and Node.js for relia- ble backend orchestration, and React.js for a responsive, user- centric frontend. For secure authentication and user manage- ment, the platform leverages JWT (JSON Web Tokens) and OAuth integration, thereby balancing robust security with seam- less access for individual programmers and collaborative teams alike.As users write or debug code, CodeWiz analyzes context, corrects errors, and proposes optimal code snippets, learning from each unique session to enhance future responsiveness and efficiency. Additional features include real-time peer-to-peer code collaboration, version management, built-in communication tools for live project discussions, and adaptive support for multi- ple programming languages. These enable teams to coordinate, code, and resolve errors together within a secure and interactive environment. CodeWiz’s personalized recommendations and di- agnostics accelerate learning for beginners, enhance productivity for experienced developers. By transforming static code editing into a holistic, adaptive, and predictive development experience, CodeWiz bridges the gap between mere syntax validation and deeper, semantic, and productivity-focused programming sup- port.The future roadmap aims to extend language support, au- tomate code documentation, and incorporate proactive project analytics, solidifying CodeWiz’s role as a breakthrough tool for the evolving demands of modern development and CS education.

Development of an Intelligent Waste Classification and Structural Reconstruction System Using a Hybrid Convolutional Autoencoder Architecture Comparing With OpenCV Keras

Authors: Aditya Ram N, Yashwanth Kumar G K, Senthil Kumar B

Abstract: In recent years, deep learning has made a substantial impact on computer vision systems, especially in image processing, feature extraction, and reconstruction. The traditional method using OpenCV with Keras-based convolutional neural networks (CNNs) has been widely employed for image analysis and classification tasks. However, such methods are often dependent on extensive manual preprocessing, require large amounts of labeled data, and involve substantial computational complexity. This paper proposes a comparative analysis of the performance of an OpenCV-Keras-based pipeline and an Autoencoder-based deep learning model for image reconstruction and representation learning. The OpenCV-Keras-based pipeline is based on a traditional supervised learning strategy, whereas the Autoencoder-based model uses an unsupervised learning strategy to learn compact representations of the input images automatically. The experimental analysis reveals that the Autoencoder-based model outperforms the OpenCV-Keras-based pipeline in terms of noise removal, feature retention, reconstruction accuracy, and computational complexity. The paper concludes that Autoencoders can serve as a more scalable and intelligent alternative to traditional OpenCV-Keras-based pipelines, especially in real-time applications. Prior studies in representation learning and reconstruction using autoencoders have demonstrated their effectiveness in noisy environments. [1], [2]

DOI:

Non-Invasive Smart Knee Support for Pain Relief and Mobility Enhancement

Authors: M.Rubina begam, K. Akhilesh, R. Dravid, G. Gobivarshan, S. Gowrinath

Abstract: A Knee pain and mobility limitations are among the most prevalent musculoskeletal disorders affecting individuals across all age groups. This paper presents a non-invasive smart knee support system designed to provide real-time pain assessment, mobility enhancement, and preventive intervention. The system integrates wearable sensors, including motion and pressure sensors, to continuously monitor knee joint activity. An intelligent algorithm processes the collected data to estimate pain levels and detect abnormal movement patterns. Based on the analysis, the system delivers immediate haptic feedback to guide corrective actions. Furthermore, the integration of Internet of Things (IoT) technology enables remote monitoring, cloud-based data storage, and long-term analysis of joint performance. Experimental results demonstrate improved user awareness, reduced strain, and enhanced mobility. The proposed system offers a proactive, user-friendly, and efficient solution for personalized knee care and rehabilitation.

Malware Detection Using Machine Learning

Authors: Asim Azhar, Richa Gupta, Mudita Saxena, Dipanshu Singh, Rahul Anjana

Abstract: This project focuses on detecting Malicious Android applications using supervised machine learning techniques. A permission based dataset is used where each application is represented by behavioral features such as requested permission. After preprocessing the dataset, machine learning algorithms including Random Forest, Decision tree and Naive Bayes are implemented using the WEKA framework in Java. The models are evaluated using 10-fold cross validation and standard performance metrics. The objective of the project is to develop an automated, accurate and safe malware detection system without executing malicious code.

DOI: https://doi.org/10.5281/zenodo.19639505

Predicting Coronary Heart Disease Using An Improved Light Gbm Model

Authors: Gomathi T, Nivedita K, Sindhu M, Soundarya P

Abstract: Coronary Heart Disease (CHD) remains one of the leading causes of mortality worldwide, necessitating early and accurate prediction methods to improve patient outcomes. This paper proposes an efficient predictive framework using an improved Light Gradient Boosting Machine (LightGBM) algorithm for the early detection of CHD. The proposed model integrates advanced preprocessing techniques, including data cleaning, normalization, and feature selection, to enhance data quality and relevance. To address class imbalance, the Synthetic Minority Oversampling Technique (SMOTE) is employed, thereby improving model robustness. Hyperparameter tuning is also performed to optimize model performance and reduce overfitting. The system is trained and evaluated using clinical datasets containing key attributes such as age, blood pressure, cholesterol levels, and lifestyle factors. Experimental results demonstrate that the improved LightGBM model achieves higher accuracy, precision, and recall compared to traditional machine learning approaches. Additionally, the model identifies significant risk factors contributing to CHD, supporting clinical decision-making. The proposed approach provides a reliable, scalable, and efficient solution for early CHD prediction, with potential applications in healthcare systems for preventive diagnosis and risk assessment. The improved LightGBM model not only enhances predictive accuracy but also reduces computational complexity, making it suitable for large-scale medical datasets. Furthermore, the interpretability of the model, achieved through feature importance analysis, enables healthcare professionals to better understand contributing risk factors and take proactive preventive measures. This approach bridges the gap between data- driven insights and clinical practice, ultimately contributing to improved patient care, early intervention, and reduced mortality associated with coronary heart disease.

AI-Powered Personalized Learning System

Authors: Shubham Prajapati, Ashutosh Kumar Mourya , Megha Mayavanshi

Abstract: The rapid advancement of Artificial Intelligence (AI) has created transformative opportunities in education. This paper presents an AI-Powered Personalized Learning System (AIPLS) that leverages machine learning, natural language processing, and adaptive algorithms to deliver customized educational experiences. The proposed system dynamically adjusts content, difficulty, pacing, and feedback based on real-time student performance analysis. Results demonstrate a 35% improvement in assessment scores and a 28% reduction in time-to-mastery compared to traditional static learning approaches.

DOI:

Heat Transfer Across A Stretching Sheet In A Viscoelastic Boundary Layer With A Non-uniform Heat Source And Viscous Dissipation

Authors: Shehzad Ali, Dr P. K. Shukla

Abstract: This work addresses visco-elastic boundary layer flow and heat transmission over a stretching sheet in the presence of viscous dissipation and a nonuniform heat source. Analytical solutions exist for highly non-linear momentum equations and heat transfer equations with confluent hypergeometric similarity. Here, two types of different heating procedures are considered: prescribed surface temperature (PST) and prescribed wall heat flux (PHF). A number of parameters, such as the visco-elastic parameter, Eckert number, Prandtl number, and non-uniform heat source/sink parameter, are analyzed in connection to the temperature distribution. Each of these parameters' effects on the wall temperature gradient and wall temperature are enumerated and described.

DOI: https://doi.org/10.5281/zenodo.19640550

Paarsh Matrimonial Website System With Intelligent Matchmaking

Authors: Nikhil Rajput, Rahul Patel

Abstract: This paper presents a web-based matrimonial website system designed for efficient partner matchmaking using structured datasets and user preference analysis. The system is implemented using Java Spring Boot, REST APIs, and MySQL database. It enables user registration, profile management, match searching, and secure communication. The study also includes system architecture, data flow diagrams, dataset usage, and research gaps. The proposed system improves matchmaking accuracy and ensures scalability and security.

Blockchain-Based Voting System Enhancing Electoral Security, Transparency, and Accessibility Through Decentralized Technology

Authors: Gourav Singh, Arjun Pataskar

Abstract: The integrity of electoral systems is fundamental to democratic governance; however, traditional voting mechanisms suffer from security vulnerabilities, lack of transparency, and accessibility constraints. This paper proposes a blockchain-based voting system leveraging distributed ledger technology to ensure secure, transparent, and tamper-resistant elections. The system integrates cryptographic techniques such as Zero-Knowledge Proofs (ZKPs) and Elliptic Curve Cryptography (ECC) within a permissioned blockchain framework using Hyperledger Fabric and Practical Byzantine Fault Tolerance (PBFT) consensus. A three-tier architecture consisting of Application, Blockchain, and Data Storage layers ensures scalability and efficiency. Security mechanisms including multi-factor authentication, end-to-end encryption, and AI-based anomaly detection mitigate potential threats such as Sybil attacks and denial-of-service attacks. Comparative analysis indicates improved security, transparency, and cost-effectiveness over traditional systems. The proposed framework demonstrates strong technical feasibility and provides a foundation for future advancements in digital electoral systems.

Ask Milo: An AI-Powered Business Consulting Platform Using Generative AI And Full-Stack Web Technologies

Authors: Deep C Patel, Prof. Ziyam Khan, Jigar M Patel

Abstract: In the age of rapidly developing technology, artificial intelligence plays a significant role in facilitating user interaction with computers. The current research introduces Askmilo, a conversational chatbot created by utilizing the technologies of MERN stack and integrating natural language processing. As the purpose of the application is to create a tool that would be able to efficiently communicate with the user via real-time conversation, it is equipped with an ability to generate context-dependent responses, maintaining chat history and thus creating the most comfortable experience for the user. The program implements such aspects as secure authentication, responsiveness, and scalability in order to provide a seamless experience when receiving multiple simultaneous requests from the users. The solution may also be easily updated with such functions as emotion detection and multimodality in the nearest future.

Multimodal Heart Disease Classification Using Ecg Signals and Clinical Data

Authors: Nivedha A. K, Sathish K, Sivadass S, Thangadurai N

Abstract: Heart disease is a major cause of mortality worldwide, making early and accurate diagnosis essential. A multimodal framework for heart disease classification is presented through the integration of electrocardiogram (ECG) signals and clinical data. ECG recordings are obtained from the PTB-XL dataset, which includes clinical attributes such as age, gender, and diagnostic labels. A Residual Neural Network (ResNet) is employed to extract discriminative features from ECG signals, while Bidirectional Encoder Representations from Transformers (BERT) is utilized to encode clinical text data and capture contextual dependencies. The extracted features are fused using a fully connected architecture to enhance classification performance. Experimental results demonstrate an accuracy of 96.77%, indicating improved performance over unimodal approaches and supporting reliable clinical decision-making.

Smart IoT-Based Wearable System For Varicose Vein Monitoring And Risk Assessment Using Hybrid Machine Learning

Authors: Saana L, Sathavarthini S, Nikitha P, Parameshwari P, Mr.Aakash M

Abstract: Varicose veins are a common vascular disorder associated with chronic venous insufficiency and may lead to severe complications if not monitored effectively. Conventional diagnostic approaches are limited to periodic clinical assessments and fail to capture continuous physiological variations. This study presents a wearable, non-invasive monitoring system for continuous varicose vein risk assessment using multi-modal sensing and intelligent data analysis. The proposed system integrates photoplethysmography (PPG), skin temperature sensing, and inertial measurement-based posture detection to acquire real-time physiological and behavioral data. A hybrid risk prediction framework combining threshold-based clinical evaluation and a Random Forest classifier is employed to improve reliability. The system is implemented on an embedded platform with wireless communication for remote monitoring. Experimental results demonstrate a classification accuracy of 94.6% with low latency and extended operational capability, indicating the effectiveness of the proposed approach for continuous and real-world monitoring of varicose vein risk.

Facial Emotion Recognition Based Smart Music Player

Authors: Drbrindhas, Ms. P.Abirami In, Mr. Ajay.R, Mr. Anbarasan.R, Mr.Rishihesh.M.M, Mr.Safwan.S, Mr.Sriram.V

Abstract: This paper presents the design and implementation of a Smart Playlist Generator using Affective Computing — a real-time, AI-driven music recommendation system that personalizes playlists based on the user's emotional state. The system integrates three core components: (1) a Facial Emotion Recognition (FER) module built on OpenCV and Convolutional Neural Networks (CNNs) that classifies emotions in real time from webcam input, (2) a Natural Language Processing (NLP) module that supports Thanglish (Tamil- English transliterated) text commands for conversational interaction, and (3) a Spotify Web API integration that maps detected emotions to audio features such as valence, energy, and tempo to generate context-aware playlists. The system achieves an emotion recognition accuracy of 87– 90%, Thanglish command interpretation accuracy exceeding 90%, and a playlist-mood alignment rate of 85–90%, with an end-to-end latency of approximately 3 seconds. The architecture leverages HTML/CSS/JavaScript for the frontend, Node.js with Express for the backend, Firebase for data persistence, and Python-based AI modules for emotion and language processing. Experimental results confirm the viability of affective computing for dynamic, personalized music delivery, and the system demonstrates significant potential for next- generation human-computer interaction in multimedia platforms. Keywords: Affective Computing, Facial Emotion Recognition, Convolutional Neural Networks, Music Recommendation System, Natural Language Processing, Spotify Web API, Thanglish Processing, Human-Computer Interaction.

DOI: https://doi.org/10.5281/zenodo.19659999

Computer Vision-Based Adaptive Traffic Signal Control With Emergency Vehicle Prioritization

Authors: Anuj Chavan, Bhagyashree Dalvi, Pooja Gajjar, Shivansh Jaiswal, Ammu J. Striney

Abstract: Urban traffic management has become increasingly difficult due to the continuous rise in the number of vehicles. Traditional traffic signal systems operate on fixed time intervals and do not adapt to real-time traffic conditions, often leading to congestion and longer waiting times. To address this issue, this paper proposes a computer vision-based adaptive traffic signal control system that dynamically adjusts signal timings based on live traffic density. The proposed system utilizes OpenCV for video processing and a YOLO-based model for real-time vehicle detection and classification. Traffic density is estimated by counting vehicles in each lane, and signal timings are assigned proportionally to improve traffic flow. In addition, the system includes a mechanism to detect emergency vehicles and provide them with immediate signal priority. The model is implemented in Python and tested in a simulated environment. Experimental results indicate improved traffic efficiency and reduced delays compared to conventional systems. Overall, the proposed solution offers a scalable and intelligent approach for modern traffic management.

Fairness-Aware Robust Handwritten Digit Recognition Using A Hybrid CNN-Boosting Framework

Authors: Prashant Kumar, Dr. Ragini Shukla

Abstract: Deep convolutional neural networks have achieved high accuracy in handwritten digit recognition; however, their reliability under adversarial perturbations, structured noise, and stylistic variation remains a significant challenge for real-world deployment. This paper presents a fairness-aware hybrid CNN-boosting framework that improves empirical robustness while reducing subgroup performance disparities. A convolutional neural network is employed as a feature extractor, and the resulting embeddings are classified using an ensemble of AdaBoost, XGBoost, and LightGBM models. Experiments on the EMNIST Digits dataset show that the proposed method attains 98.45% accuracy on clean data, outperforming a standalone CNN baseline (96.85%). Under Fast Gradient Sign Method (FGSM) attack with ε = 0.1, the ensemble achieves better retention stability than the baseline (0.866 vs. 0.848). The framework also demonstrates strong resilience to salt-and-pepper noise and 20% pixel occlusion. Fairness analysis across stroke-thickness subgroups indicates that loss reweighting reduces performance disparities without sacrificing overall accuracy. Cross-domain evaluation, however, reveals that distribution shift remains a persistent challenge despite gains in perturbation robustness. Overall, the results suggest that combining ensemble diversity with fairness-aware optimization offers a practical and scalable approach to building more robust and equitable handwritten digit recognition systems.

DOI: https://doi.org/10.5281/zenodo.19660690

Comprehensive Study of Software Testing Techniques and Their Impact on Web Application Quality

Authors: Kundan Kumar Singh, Prof. Mr..Prashant Kothari

Abstract: The growing complexity of web applications in domains such as e-commerce, healthcare, financial services, and government portals has significantly raised the bar for software quality. Delivering a web application that is functionally correct, performant, secure, and accessible across diverse devices and browsers demands a rigorous, structured approach to software testing. This paper presents a comprehensive study of software testing techniques and their measurable impact on web application quality. It examines both functional and non-functional testing strategies, explores formal test case design methodologies including Equivalence Partitioning, Boundary Value Analysis, and Decision Table Testing and analyses the role of test automation using Selenium WebDriver in improving test coverage and execution efficiency. The paper further investigates defect tracking and lifecycle management using Jira, and discusses real-world testing observations drawn from the testing of a live Government Scheme Management System. Findings indicate that the systematic application of diversified testing techniques, when integrated within an Agile Software Testing Life Cycle, results in significant reductions in post-release defect rates, improved user satisfaction, and enhanced system reliability. The study concludes by proposing a structured Testing Quality Framework suitable for adoption in modern web development environments.

A Hybrid Multi-Task Deep Learning Framework For Brain Tumor Classification And Segmentation Using ResNet-U-Net Architecture

Authors: Dr.K. Ravikiran, Guruvannagari Rishi, Veerabathini Sai Kumar, Uliravula Akhil

Abstract: Accurate identification and definition of brain tu-mors is a key element in effective diagnosis and treatment planning which at the same time is very time consuming and subject to variable results with manual analysis of medical images. We present a hybrid modality multi task deep learning framework for at the same time classifying and segmenting brain tumors which uses MRI and CT scans. We have put together a ResNet-50 encoder with a U-Net based decoder which enables joint learning of spatial and semantic features in a single architecture. Also we have used a dual branch design which at the same time produces pixel level tumor segmentation and at the same time determines tumor types which include Glioma, Meningioma, Pituitary, and Normal cases. To solve for the issues of different imaging modalities we put forth a unified preprocessing pipeline which allows the model to learn modality invariant features. We report we see a classification accuracy of 97.40% and a Dice similarity coefficient of 0.843 with our put forth framework also at the same time reporting that we are able to perform efficient real time inference on CPU based systems. Also we see that multi task learning does in fact improve diagnostic performance which in turn shows off the put forth system’s value as a practical tool for clinical decision support.

HeirCloud: A Secure Framework For Posthumous Asset Management Using Heir Access Code-Based Encryption

Authors: Praveenkumar M, SathishRaj M, Thamizharasan S, Sanjai U, Dr. M. Rajesh Babu

Abstract: Posthumous data includes an individual’s digital information such as social media accounts, emails, financial records, personal documents, and legal assets. Managing this data after death presents challenges including lack of secure access mechanisms, risk of unauthorized use, and absence of reliable delivery systems. Existing approaches, such as static instructions or platform-specific legacy features, often fail to provide strong encryption, fine-grained access control, and automated data sharing. To address these limitations, this paper proposes a secure cloud-based posthumous data management system based on the Heir Access Code-based Encryption (HACE) scheme integrated with dynamic access control. Sensitive data is encrypted using unique keys, while heir-specific access codes are mapped to the encrypted data, enabling only authorized heirs to access assigned information. This ensures selective data sharing without exposing the entire dataset. The system further incorporates dynamic policy management, allowing users to update access permissions over time, along with activity monitoring and verification mechanisms to ensure controlled data release. By combining encryption, access code-based authorization, and automated verification, the proposed system provides a reliable and scalable solution for managing posthumous digital assets.

DOI:

Performance Analysis Of PID Controller For DC Motor Speed Regulation

Authors: Veena Vanamane, Shreya S, Karthik, Ajay BK, Praneeth

Abstract: The evaluation of a proportional-integral-derivative (PID) controller's performance for efficient DC motor speed control. Based on the DC motor's electrical and mechanical properties, a mathematical model is created, and a closed-loop control system is created. In order to attain better dynamic performance, such as shorter rising times, less overshoot, and low steady-state error, the PID controller's parameters are adjusted. MATLAB/Simulink is used for simulation in order to assess the reaction of the system under various operating situations. The findings show that, in comparison to open-loop operation, the PID controller greatly improves the precision and stability of motor speed control. The study shows that a well-tuned PID controller is a dependable and effective way to regulate DC motor speed in industrial settings.