Feature Selection in Intrusion Detection Grey Wolf Optimizer Devi E M Roopa*, Suganthe R C** *Assistant Professor, Kongu Engineering College, Erode, Tamil Nadu, India **Professor, Kongu Engineering College, Erode, Tamil Nadu, India. roopasen5@gmail.com Online published on 23 March, 2017. Abstract With the growth of network based services sharing sensitive information through internet has been increased and hence network security is at risk. The intruders and attacks are ever growing and become a serious problem to detect. The labeling of audit data manually consumes more time, expensive and tedious. Since the ability to identify important inputs that can reduce size, training time and improve accurate results, it is critical to mark the important feature of network traffic for intrusion detection using classifier that would achieve a higher performance. In this paper we have used gray wolf optimizer a swarm based optimization method to search the feature space to find optimal feature subset that improves classification accuracy. At first the grey wolf optimizer uses filter-based principles to find solutions with minor redundancy that are described by mutual information. At the later stage optimization wrapper approach is employed for guiding classifier performance. The performance of grey wolf optimizer is measured and compared against several other metaheuristic algorithms with the help of NSL KDD Dataset. Top Keywords Classification, Feature Selection, Intrusion Detection, Naïve Bayes, Support Vector Machine (SVM). Top |