Comparative analysis of gravitational search algorithm and k-means clustering algorithm for intrusion detection system

Abstract

Intrusion Detection System (IDS) is an active defense technology. Many clustering algorithms are used to improve the performance of accuracy and hit rate and reduce False Alarm Rate (FAR). Conventional k-Means is the most popular clustering algorithms due to its simplicity and efficiency. However, its performance is highly dependent on the initial centroid and may trap in local optima. In recent years, heuristic algorithms have been applied to solve clustering problems. Gravitational Search Algorithm which is one of the newest swarm intelligent provides a prototype classifier to address the classification of instances in multiclass datasets. This paper used KDD Cup 1999 dataset to evaluate the performance of the baseline k-Means and GSA-based classifier in terms of accuracy, FAR and hit rate. The results show that GSA has a capability in order to improve the performance of the system.