Feature Subset Selection Using Genetic Algorithm for Intrusion Detection System

Author: Behjat Amir Rajabi   Vatankhah Najmeh   Mustapha Aida  

Publisher: American Scientific Publishers

ISSN: 1936-6612

Source: Advanced Science Letters, Vol.20, Iss.1, 2014-01, pp. : 235-238

Disclaimer: Any content in publications that violate the sovereignty, the constitution or regulations of the PRC is not accepted or approved by CNPIEC.

Previous Menu Next

Abstract

KDD 99 intrusion detection datasets, which are based on DARPA 98, is a labeled dataset studied in the field of intrusion detection. The respected KDD dataset contains numerous features, in which some of them are irrelevant or less effective to detect the attacks. This study is set to improve the classification of intrusions by means of selecting significant features. Binary Genetic Algorithm (BGA) is proposed for feature selection in order to decrease the number of unrelated features. The selected features are then become the input for the classification task using a standard Multi-layer Perceptron (MLP) classifier. The results achieved show very high classification accuracy and low false positive rate with the lowest CPU time.