Intrusion Detection and Prevention Using CNN-LSTM

Authors- T Sai Harshitha, V. Sreenidhi, Sk. Parveen, P Tejaswini, Assistant Professor Yerininti Venkata Narayana

Abstract-The rapidly evolving landscape of digital interactions, the persistent threat of intrusion attacks poses significant challenges to the security and stability of computer networks. The effects of such attacks are multifaceted, ranging from service disruptions and data breaches to the compromise of sensitive information and financial losses. The project, titled “Detection and Prevention of Intrusion Using CNN-LSTM” introduces an advanced method for network security. By combining Convolutional Neural Networks (CNN) and Long Short-Term Memory (LSTM) networks, the model can effectively detect and prevent intrusions. The study involves training the model with diverse datasets containing both normal and malicious network behaviors. The CNN part focuses on spotting spatial patterns, essential for identifying specific types of attacks, while the LSTM part captures time-related patterns in network traffic. The integrated model works in real-time, keeping a constant eye on network activities and proactively blocking suspicious actions. Deep learning models like this can adapt to new cyber threats over time. The project focuses on three attacks namely: DoS (Denial of Service), Malware, Portscanning. The research includes experiments on popular datasets, and its practical implementation involves taking actions such as implementing quarantine system and blocking suspicious ports based on the model’s predictions.

DOI: /10.61463/ijset.vol.12.issue2.125