A Comparative Analysis of Intrusion Detection Systems using Feature Selection and Classification

Abstract

Intrusion Detection Systems (IDSs) are playing major role in the process of detecting the various kinds of attacks in networks. Distance between the nodes in network is used to make decision over the nodes and network by identifying the inlier and outlier. The outlier detection is an important task in clustering and classification processes. The distance measurement formulae are used to measure the distance between the nodes in network. By applying the measurement formulae are important in IDSs. In this paper, we analyse the various IDSs which are applying the distance measurement formulae, outlier detection, feature selection and classification. The various distance measurement formulae are discussed in detail and the various outlier detection algorithms, feature selection algorithms and classification algorithms. The soft computing techniques based classification algorithms are also analysed in this paper. Moreover, we have compared the different combinations of IDSs based on the experimental results. Finally, we have proposed a suggestion to achieve better classification accuracy, low false alarm rate and better life time than the existing works.