A review of support vector machine-based intrusion detection system for wireless sensor network with different kernel functions

Abstract

Wireless sensor network (WSN) is among the popular communication technology which capable of self-configured and infrastructure-less wireless networks to monitor physical or environmental conditions. WSN also is the most standard services employed in commercial and industrial applications, because of its technical development in a processor, communication, and low-power usage of embedded computing devices. However, WSN is vulnerable due to the dynamic nature of wireless network. One of the best solutions to mitigate the risk is implementing Intrusion Detection System (IDS) to the network. Numerous researches were done to improve the efficiency of WSN-IDS because attacks in networks has been evolved due to the rapid growth of technology. Support Vector Machine (SVM) is one of the best algorithms for the enhancement of WSN-IDS. Nevertheless, the efficiency of classification in SVM is based on the kernel function used. Since dynamic environment of WSN consist of nonlinear data, linear classification of SVM has limitations in maximizing its margin during the classification. It is important to have the best kernel in classifying nonlinear data as the main goal of SVM to maximize the margin in the feature space during classification. In this research, kernel function of SVM such as Linear, RBF, Polynomial and Sigmoid were used separately in data classification. In addition, a modified version of KDD’99, NSL-KDD was used for the experiment of this research. Performance evaluation was made based on the experimental result obtained. Finally, this research found out that RBF kernel provides the best classification result with 91% accuracy.