Design and development of intelligent agent based intrusion detection system using machine learning methods

Abstract

Cybersecurity is a critical aspect of information technology to safeguard data across the network. Sophisticated attacks are increasing because of technological advances. Detecting these attacks is a significant issue in today s digital era. Intrusion Detection Systems play a vital role in ensuring a secure network by identifying malicious threats. To handle the changing nature of attacks and to improve efficiency, an enhanced machine learning based Intrusion Detection System is used by integrating an newlineintelligent feature-selecting agent with an SHiP vector machine for classification. The approach consists of two modules: IV-RFE feature selection agent, which selects the relevant features, and SHiP Vector Machine algorithm, which detects the intrusions effectively. The main aim of the proposed research is to detect diverse attacks by consistently maintaining a high performance across diverse attacks. The proposed model is tested against attacks, namely, Reconnaissance, Analysis and DoS from the newlineUNSWNB15 dataset. To test the robustness of the model, the proposed model is validated against a realtime dataset named PSD-23 containing DoS attacks. The comprehensive results highlight the enhanced effectiveness of the proposed model in comparison with existing approaches.