ICSEMT 2025 Proceeding

Proceeding of the Conference

A Comparative Study on the Erosive Wear Characteristics of HDPE/Tantalum Iron Using Titanium/Rutile Sludge as Abrasive

Authors- Abhinand G, Jayashankar P, Rojan Antony, Sanjo Paul Kalluvilayathil, V.R. Rajeev, Ajith G. Nair and Muhammed Arif M

Abstract- – This paper aimed to study the erosive wear characteristics of high density polyethylene (HDPE) and Tantalum iron using Titanium/Rutile sludge as an abrasive material. The sludge is obtained from Travancore Titanium Products Ltd(TTPL) which is a waste material with 6 to 6.5 hardness on mohs scale. An inhouse built erosive wear test rig will be used to simulate the erosive wear characteristics of HDPE/ Tantalum iron using Titanium/Rutile as abrasive. The process parameters, including specimen velocity, slurry concentration, impact angle, and time are varied within fixed ranges. From the studies it reached to the conclusion that erosive wear is more on Tantalum iron when compared to HDPE and time has more contribution to erosive wear loss.

Analysis of Authentication Protocols with Finite Field Cryptography

Authors- Arood Ahmad Dar, Sarabjit Kaur

Abstract- – The research investigates the design, analysis, and implementation of privacy-preserving authentication protocols leveraging techniques from finite field cryptography. Authentication is a fundamental security mechanism in modern communication systems, ensuring that entities can securely verify each other’s identities. However, traditional authentication protocols may compromise users’ privacy by revealing sensitive information during the authentication process. This research aims to develop novel authentication protocols that provide strong security guarantees while preserving user privacy through the use of finite field cryptography. The research explores theoretical foundations, protocol design, security analysis, and practical considerations, contributing to advancements in privacy-preserving authentication mechanisms.

Finite Field Optimizations for Low-Latency Communications

Authors- Arood Ahmad Dar, Dr.Sarabjit Kaur

Abstract- – The exponential growth of secure communication technologies, including 5G/6G networks, IoT, and autonomous systems, has necessitated cryptographic protocols that ensure security while maintaining low latency. Finite fields (GF(pn)\text{GF}(p^n)GF(pn)) underpin many cryptographic algorithms and error-correction systems. However, their inherent computational complexity often hinders real-time performance. This paper explores mathematical optimizations and hardware implementations for finite field operations, focusing on latency-critical applications. The proposed methods achieve significant reductions in computation time for field arithmetic, efficient basis representation, and hardware acceleration, validated through simulations and real-world testing in next- generation communication systems.

Sleep Health Monitoring through Arduino UNO & EXG

Authors- Shashank Kumar, Nickey Ray, Dr. Manoj Kumar Pandey, Dr. Sakshi Kathuria

Abstract- – This article outlines the minimum requirements for record- ing standard and sleep EXG. The Epilepsy Guidelines Work- ing Group, in collaboration with the International League Against Epilepsy and the International Federation of Clini- cal Neurophysiology, has developed guidelines for clinical practice guidelines related to epilepsy. All studies were eval- uated using PRISMA and GRADE evidence rating, with QUADAS-2 assessing the risk of bias in technical and meth- odological review studies related to sleep induction meth- ods. We will use a modified version of Delphi for the is-sue, where we couldn’t find enough credible public evidence to establish consensus on expertise. The process was aided by the GRADE system to generate recommendations. We em- ployed evidence that was either of low or moderate value. Our team developed 16 scenarios based on consensus to es- tablish minimum standards for recording routine and sleep EXG. Technical specifications, recording duration, induce- ment to sleep time and provocative methods are part of the recommendation.

Comparative analysis of Machine Learning Applications in IoT for Effective ERP Management

Authors- Research Scholar Ms. Kanika Yadav, Assistant Professor Dr. Bijendra Singh

Abstract- – Integrating ML and IoT with ERP systems improves data-driven decision-making. This study simulates an IoT-based data collection system that records temperatures, humidity, and inventory levels in a CSV file. Environmental circumstances predict inventory levels using a binary classification model. Training the model using temperature and humidity and a Random Forest classifier. A 100-simulated IoT reading dataset is processed, scaled, and separated into training and test sets. F1-score, accuracy, precision, and recall assess model performance. ROC curve, confusion matrix, and error curve visualizations indicate model performance. Machine learning can anticipate stock levels using real-time IoT data to improve ERP inventory management. This strategy improves supply networks using predictive analytics to improve operational efficiency and decision-making. This study informs ERP scalable ML-IoT integration research. Integrating ML and IoT with ERP systems improves data-driven decision-making. This study simulates an IoT-based data collection system that records temperatures, humidity, and inventory levels in a CSV file. Environmental circumstances predict inventory levels using a binary classification model. Training the model using temperature and humidity and a Random Forest classifier. A 100-simulated IoT reading dataset is processed, scaled, and separated into training and test sets. F1-score, accuracy, precision, and recall assess model performance. ROC curve, confusion matrix, and error curve visualizations indicate model performance. Machine learning can anticipate stock levels using real-time IoT data to improve ERP inventory management. This strategy improves supply networks using predictive analytics to improve operational efficiency and decision-making. This study informs ERP scalable ML-IoT integration research.

A Comprehensive Analysis of Hybrid Machine Learning Models for Social Media Threat Detection and Forecasting

Authors- Research Scholar Ms. Rashmi Tiwari, Professor Dr. Gaurav Aggarwal

Abstract- – The rapid proliferation of social media platforms has fostered unprecedented connectivity while also creating new avenues for malicious activities such as cyberbullying, misinformation, and threats. To address these challenges, this paper explores hybrid machine learning models designed to enhance the detection and prediction of threats on social media. By integrating diverse algorithmic paradigms, hybrid models leverage the strengths of different approaches, offering a robust and scalable solution. This study reviews existing literature, identifies key challenges, and presents a roadmap for developing effective hybrid frameworks for social media threat analysis.

Impact of IEC on the Association between Socio-Demographic Factors and Knowledge Regarding Type II Diabetes Mellitus Prevention in a Rural Population of Kathua, Jammu & Kashmir

Authors- Shally Sharma, Dr. Khemchand

Abstract- – Background: Type II Diabetes Mellitus (T2DM) remains a significant public health issue, especially in rural regions where awareness of preventive measures is often inadequate. Socio-demographic factors play a crucial role in determining individuals’ knowledge and their capacity to implement preventive strategies. This study investigates the relationship between socio-demographic variables and knowledge about T2DM prevention, assessing changes before and after an IEC-based educational intervention. Methods: A quantitative cross-sectional study employing a pre-test-post-test design was carried out among 400 rural participants in Kathua, J&K. Participants’ knowledge of T2DM prevention was evaluated using a structured questionnaire. The relationship between socio-demographic factors and knowledge levels was examined through chi-square analysis. Results: Chi-square analysis revealed significant associations between socio-demographic factors (age, gender, education, marital status, occupation, family income, social habits, and type of family) and knowledge levels regarding T2DM prevention, both in the pre-test and post-test phases (p < 0.05). These associations highlight how different socio-demographic variables influence participants' knowledge about T2DM prevention. Conclusion: Socio-demographic factors are significantly associated with knowledge levels regarding T2DM prevention. The IEC intervention effectively modified these associations, highlighting the need for targeted education strategies in rural public health programs.

Hybrid IOT Based Model for Improving E-Healthcare System

Authors- Ms. Niyati Jain, Dr. Kavita Mittal

Abstract- – Background: This study evaluates an IoT-enabled healthcare system for predicting diabetes patient outcomes using LSTM network. The proposed LSTM-based model outperforms traditional methods in key metrics such as accuracy, precision, recall, and F1-score. The study uses a dataset of 768 records of Pima Indian female patients aged 21 and above, with nine features: pregnancies, glucose level, blood pressure, skin thickness, insulin, BMI, diabetes pedigree function, age, and outcome. The LSTM network is designed to address the temporal dependencies and complex relationships inherent in the dataset. The study compares the performance of the LSTM-based system with traditional models through the construction of confusion matrices. These matrices are used to compute and analyze F1 scores, recall values, precision, and overall accuracy of the models. The proposed system exhibits significantly improved performance across all metrics, underscoring its efficacy in accurately predicting diabetes outcomes. The accuracy achieved by the LSTM model surpasses that of traditional models, while the error rate is notably lower, emphasizing its reliability and robustness. A comparative analysis with prior research highlights the advantages and potential limitations of the proposed LSTM-based IoT healthcare system. While it excels in handling sequential and time-dependent data, it requires a substantial amount of computational resources and training time. Despite these limitations, the improved prediction outcomes justify its implementation in real-world IoT healthcare systems, where accuracy and reliability are paramount.

Hybrid Encrypted Socket Towards Security in Cloud Environment

Authors- Ms. Radhika Garg, Dr. Kavita Mittal

Abstract- – The proposed encryption method offers significant benefits over traditional cryptographic approaches, including improved performance and security in time consumption, error rate, and resistance to various types of attacks. It consistently outperforms DES, RSA, AES, and DNA encryption in terms of time consumption, error rate, along with security analysis. The method’s reduced packet size and compression techniques enhance processing speed and decrease latency, reducing the likelihood of errors during data transfer. In terms of error rates, proposed method shows a significant reduction compared to other methods, minimizing packet size and transmission time, thereby reducing the likelihood of errors during data transfer. It also excels in security analysis, with lower susceptibility to attacks such as, brute force, denial-of-service, and access violations. It demonstrates superior resilience against Man-in-the-Middle attacks and brute force attempts, achieving fewer successful breaches than DES, RSA, AES, and DNA encryption methods. The advent of CC has made it possible for individuals to test out novel concepts, such as managing digital resources and content calls. Current research has identified potential security models such as RSA, AES, DES, and DNA protection for safeguarding data stored on the cloud. However, the effectiveness of clouds has only been the subject of limited studies. Researchers have conducted extensive investigations to enhance the cloud’s performance and security level, with the goal of increasing safety levels without compromising operational efficacy. Cloud computing provides platforms for innovative practices, such as building digital resources and content management. Existing security models, such as mechanism, DNA security, and several security protocols, have been used to secure cloud-based contents. However, limited research has focused on cloud performance, and the proposed solution should improve security without affecting or decreasing performance.

Analyzing the security during Big Data Transmission Over the Network

Authors- Ms. Sugandha, Dr. Kavita Mittal

Abstract- – This research aims to enhance the security and performance of big data transmission across networks by proposing a secure clustering approach that integrates encryption mechanisms. The study addresses four primary objectives: analyzing security risks to structured and unstructured data through a literature review, evaluating clustering and encryption techniques to improve big data security, developing a robust model for secure data transmission, and conducting a comparative analysis of the proposed model against existing methods. The proposed framework leverages a combination of clustering, and encryption to safeguard data while ensuring efficient transmission. The data flow diagram illustrates a multi-stage process, beginning with parallel operations to extract relevant datasets and keywords. The frequency of these terms is determined using MapReduce and grouping methods. Enhancements in K-means clustering involve iteratively linking data entries to the nearest centers, recalculating centroids, and optimizing cluster configurations until stability is achieved. The secure framework is built upon a layered architecture comprising client-side validation, server-side storage, database management, map reduction, and encryption for content protection. The implementation involves the use of a Hadoop-based environment, integrating tools such as Spark and Hive, along with Python scripts for data processing. The comparative analysis highlights the efficiency of Spark’s in-memory execution, which significantly outperforms MapReduce in processing speeds. Hive, with its concise query capabilities, is shown to be more efficient than Pig in handling large datasets. Python scripts and Cloudera integration further streamline data processing, ensuring scalability and reliability. Experimental results underscore the performance advantages of the proposed model. Accuracy metrics, are derived from confusion matrices for both optimization and non-optimization datasets. The proposed clustering and encryption techniques achieved an overall accuracy of 93.4% with optimized datasets, compared to 92.2% for non-optimized datasets. The study also demonstrates the effectiveness of the K-means algorithm in identifying clusters, with an accuracy of 86.05% for raw data.

Cloud Migration Technique: A Review

Authors- Prevesh Kumar Bishnoi, Dr. Dharminder Kumar, Dr. Prateek Bhanti

Abstract- – By means of a process of transformation, companies shift their data, programs, and workload from on-site infrastructure to cloud environments. Combined with other cloud migration strategies, this study looks at rehosting, replatforming, refactoring, hybrid cloud solutions combined with their benefits and drawbacks. Analyzed are how generative artificial intelligence, machine learning, and optimization techniques could lower downtime and consequently maximize resource use, thereby enhancing migration efficiency. Investigated in order to solve data flow issues are virtual machine (VM) migration strategies—pre-copy and compression-based approaches. Among the underlined issues are security ones; GDPR and HIPAA compliance; regulatory observance; access control; encryption. Furthermore, addressed are energy-efficient migration strategies meant to reduce running expenses and power consumption. The book underlines the need of industry case studies and practical applications in grasping the complexity of migration. This paper gives a whole picture of cloud migration strategies by tackling scalability, flexibility, and cost economy. Future studies should concentrate on improving security measures and on creating artificial intelligence-driven automated solutions to support cloud migration procedures even more. The results provide companies looking for sensible solutions for cloud migration with interesting analysis.