Intrusion Detection and Classification using Decision Tree Based Key Feature Selection Classifiers

Abstract

Feature selection method applied on an intrusion dataset is used to classify the intrusion data as normal or intrusive.We have made an attempt to detect and classify the intrusion data using rank-based feature selection classifiers.A set of redundant features having null rank value are eliminated then the performance evaluation using various feature selection algorithms are done to determine the behavior of attributes.We can distinguish the key features which plays an important role for detecting intrusions.There are 41 features in the dataset, out of which some features play significant role in detecting the intrusions and others do not contribute in the detection process.We have applied different feature selection techniques to select the predominant features that are actually effective in detecting intrusions.