AI- Driven Identification of Irregular Network Behaviour

Abstract

Ensuring network security has become a crucial concern due to the explosive proliferation of computer networks and internet-based services. Firewalls and signature-based intrusion detection systems are examples of traditional security measures that are frequently inadequate to identify changing and unidentified attack methods. The design and implementation of an AI-Driven Identification of Irregular Network Behaviour system, which uses machine learning techniques to identify anomalies in network data, are presented in this study. In order to categorise traffic as normal or irregular, the suggested system examines network flow characteristics taken from benchmark datasets like NSLKDD and CICIDS. For precise detection, a Random Forest-based model is used; for real-time prediction and visualisation, the system is coupled with a Flask-based web application.