AI-based Intrusion Detection System (IDS) using Behaviour Analysis

Abstract

A rise in cyber threats, such as Denial of Service (DoS) attacks, has highlighted the limitations of traditional Intrusion Detection Systems (IDS) in Malaysia’s cyberspace. Existing IDS face difficulties in detecting new threats due to reliance on signature databases and high false positive rates. To address these issues, an AI-based IDS utilizing Long Short-Term Memory (LSTM) networks and behavior analysis techniques is proposed. LSTM improves anomaly detection by learning traffic patterns over time, while behavior analysis enhances accuracy by focusing on deviations from normal network behavior. The tool consists of three modules: monitoring, anomaly detection and user interface. The AI-based IDS tool aims to provide real-time threat detection for individuals and organizations. The expected outcome is a reliable IDS with improved accuracy and reduced false positives, promoting safer network environments.