Clustering Algorithms for Intrusion Detection
K. S. Anil Kumar, Anitha Mary M. O. Chacko · 2016
An intrusion detection system is an intelligent system developed to identify and counteract intrusive efforts. Clustering algorithms are used in intrusion detection systems for separating normal activities from abnormal activities. The selection of an efficient clustering technique is a highly challenging job. This article analyzes the different clustering techniques that can be used in the preprocessing stage of intrusion detection systems. This paper compares the performances of three clustering techniques on DARPA dataset- K-means clustering, K-medoid clustering and Improved K-means clustering technique with optimum cluster centroid initialization algorithm. The best recognition results with an accuracy of 98.58% was achieved using the Improved K-means algorithm.