Daniel of service attack detection using classification techniques in wsns

Abstract

Wireless Sensor Networks (WSNs) are comprised of a large number of sensor nodes that are low in cost and smaller in size. The sensor nodes are usually placed in open areas and used in many applications. The nature of WSNs makes it threatened by many security attacks, one of them is the Denial of Service (DoS) attack which is defined as any activities that prevent the network to perform its expected functions. Intrusion Detection System (IDS) is a mechanism used to detect the malicious nodes in the network. In this paper, classification techniques are used as a tool to detect intruder node. These techniques are Naives Bayesian, Support Vector Machine (SVM), Random Forest and J48. Four types of DoS attacks are considered in this study, they are: Blackhole, Grayhole, Flooding and Scheduling attacks. The detection performance evaluation is measured by different metrics such as True Positive Rate (TP), Precision (P), Recall, False Positive Rate (FP) and ROC area. A specialized dataset for WSNs is used as an input file. Using WEKA data mining tool. The results show that the SVM classifier outperforms the other classifiers with high detection rate of 96.7%.