A Novel Approach for Intrusion Detection Using Swarm Intelligence

This paper presents Architecture (IHDAIDS) for Intrusion Detection in wired networks which used Hybrid Intelligent Intrusion Detection Techniques making use of Swarm Intelligence algorithms. The architecture proposed in this paper is Intelligent, Hybrid, Distributed and Adaptive which makes real-time and dependable decisions regarding the intrusions; the Intelligent and Adaptive nature of the engine minimizes the rate of false positives and increases the accuracy, also reduces human intervention. This architecture is also concentrated on the data collection module as the quality of the intrusion detection depends on the data provided to the intrusion detection engine. We used a Hybrid Swarm Intelligence algorithm PSOACO2 (Particle Swarm Optimization/Ant Colony Optimization) for intrusion detection and compared the results with SVM (Support Vector Machine). For experiments we considered KDDCUP'99 Data which is widely used by intrusion detection researchers as a standard.