Revolutionizing Network Security: Integrating Convolutional Neural Networks for Enhanced Real-Time Intrusion Detection and Automated Attack Classification

Abstract

This work describes the development of a mixture of both the machine learning and deep learning based Intrusion Detection System (IDS) for the purpose of improving the real time identification of threats on a network. Out of the traditional Machine Learning models Decision Trees, Random Forests, and Logistic Regression are used but have the drawback of not being able to identify evolving attacks so the proposed model has CNN incorporated into it. Thus, the CNN-based IDS enhances detection performance and versatility with respect to diverse threats, including DDoS, DoS, and Port Scans. The system provides capabilities such as the live log stream, integrated attack categorization, and use of Streamlit for easy control over the detection flow. Performance evaluation on three different network datasets shows that the developed system offers an efficient solution for recognizing new and existing attacks with improved precision. The IDS offers an optimal security solution which can effectively work and could be scaled up as necessary.