Feature Subset Evaluation and Classification using Naive Bayes Classifier

Feature Reduction is the reduction of features. Most of the intrusion detection approaches focused on feature selection issues such as irrelevancy, redundancy and length of detection process. These issues will degrade the performance of system. The performance of the system is improved by three feature selection methods involving correlation based feature selection, Gain Ratio and Information Gain. The threshold based Naive feature reduction algorithm is used to reduce the features. The reduced features are further classified by Naive Bayes classifier to produce best performance to design Intrusion Detection System. Index Terms – Feature selection, classification, Feature reduction, Intrusion detection.