Anomaly based Intrusion Detection using Hybrid Learning Approach of combining k-Medoids Clustering and Naive Bayes Classification

The role of Intrusion Detection System (IDS) has been inevitable in the area of Information and Network Security - specially for building a good network defense infrastructure. Anomaly based intrusion detection technique is one of the building blocks of such a foundation. In this paper, the attempt has been made to apply hybrid learning approach by combining k-Medoids based clustering technique followed by Naive Bayes classification technique. Because of the fact that k-Medoids clustering techniques represent the real world scenario of data distribution, the proposed enhanced approach will group the whole data into corresponding clusters more accurately than k-Means such that it results in a better classification. An experiment is carried out in order to evaluate performance, accuracy, detection rate and false positive rate of the classification scheme. Results and analyses show that the proposed approach has enhanced the detection rate with minimum false positive rates.