An Intrusion Detection Model Based on Fuzzy C-means Algorithm

Abstract

Massive researches indicated that intrusion detection model created by combining unsupervised learning and supervised learning algorithm have shown better detection performance. In the process of intrusion detection, huge size of the data and unbalance of normal data and intrusion data were inevitable obstacles. In order to solve those problems, fuzzy c-means (FCM) algorithm and KNN algorithm were applied to reconstruct feature vectors based on central points and train classifier, respectively. The experiment results on KDD-Cup 99 dataset show that this algorithm can achieve higher accuracy than other similar ones on unbalanced distribution data.