Building an Effective Approach toward Intrusion Detection Using Ensemble Feature Selection

The duplicate and insignificant features present in the data set to cause a long-term problem in the classification of network or web traffic. The insignificant features not only decrease the classification performance but also prevent a classifier from making accurate decisions, exclusively when substantial volumes of data are managed. In this article, the author introduced an ensemble feature selection (EFS) technique, where multiple homogeneous feature selection (FS) methods are combined to choose the optimal subset of relevant and non-redundant features. An intrusion detection system, named support vector machine-based IDS (SVM-IDS), is prompted using the feature selected by the proposed method. The SVM-IDS performance is evaluated using two benchmark datasets of intrusion detection, including KDD Cup 99 and NSL-KDD. Our proposed method provided more significant features for SVM-IDS and compared with the other state-of-the-art methods. The experimental results demonstrate that proposed method achieves a maximum accuracy as 98.95% in KDD Cup 99 data set and 98.12% in the NSL-KDD data set.