An IWD-based feature selection method for intrusion detection system

Intrusion detection system (IDS) is an essential cyber security tool which is used to detect abnormal activity on a network or a host. A general approach towards designing IDS models is to use classifiers as detection units. But a large feature space including noisy, redundant and irrelevant features often leads to low detection and high misclassification rates by the classifier. To address this drawback, the process of selecting most relevant key features for classification is highly important. The objective of this work is to optimize the process of feature selection in a way that improves the accuracy of the classifier. This paper presents an IDS model wherein an intelligent water drops (IWD) algorithm-based feature selection method is proposed. This method uses the IWD algorithm, a nature-inspired optimization algorithm for the feature subset selection along with support vector machine as a classifier for evaluation of the features selected. The experiments are conducted using KDD CUP'99 dataset, and the performance is compared with earlier designs. The experimental results show that the proposed model performs better in terms of higher detection rate, low false alarm rate and improved accuracy than the existing approaches.