Performance analysis of machine learning algorithms for intrusion detection in MANETs

Mobile Ad-hoc network (MANET) has become an important technology in recent years and the corresponding security problems are getting more and more attention. In this paper, we apply seven well-known machine learning algorithms to detect intrusions in MANETs. We have generated training data under various simulation parameters. We also propose a new measure method which uses five new features to represent the network traffic. The analysis results show that the multilayer perceptron, logistic regression and Support Vector Machine (SVM) have the best performance and the logistic regression and SVM also get very little time to train the classification model. Copyright © 2013 Inderscience Enterprises Ltd.