A Hybrid Bat Based Feature Selection Approach for Intrusion Detection

Abstract

Intrusion detection Systems (IDS) are used for detecting malicious and abnormal behaviors, but they suffer from many issues like high resource consumption, high false alarm rate and many others. In this paper, we present a new algorithm to improve intrusion detection and reduce resource consumption. The proposed HBA-SVM IDS combines a hybrid Bat meta-heuristic Algorithm with a support vector machine (SVM) classifier for simultaneous feature and optimal SVM parameters selection, to reduce data dimensionality and to improve IDS detection. To evaluate our system, we used the NSL-KDD dataset and compare against a standard SVM and a PSO-SVM algorithm. Compared to these algorithms experimental result show that our system reduces the number of features needed for intrusion detection by 62% and achieves higher detection rate and lower false alarm rate.