Least Square Support Vector Machine based IDS, using feature selection algorithm

Abstract: When the different users on the Internet access similar content which may be redundant or irrelevant data features which causes problems in network traffic classification. This retards the network traffic classification process and prevents to make accurate and optimal decisions when are dealing with big data. In this paper A hybrid feature selection algorithm is used for optimal feature classification and these mutual information based algorithms can handle both linearly and nonlinearly dependent data features. The results will be evaluated during network intrusion detection. The Least Square Support Vector Machine based IDS (LSSVM-IDS) which is an Intrusion Detection system and is developed using features of feature selection algorithm and its performance is evaluated using the data sets provided by KDD Cup 99 data sets.