WSEAS Transactions on Communications
Print ISSN: 1109-2742, E-ISSN: 2224-2864
Volume 13, 2014
Cross Layer Intrusion Detection System of Mobile Ad Hoc Networks Using Feature Selection Approach
Authors: ,
Abstract: Existing intrusion detection system of Mobile ad hoc networks (MANET) examines all the features of network audit data. Some of the features of collected data may be redundant or irrelevant to the detection process. So it is vital to select the important features to increase the detection accuracy. This paper focuses on implementing two feature selection methods namely, Rough Set Theory (RST) and genetic algorithm (GA) combined with Support Vector Machines (SVM). Also the proposed system uses cross layer features instead of single layer features to maximize the performance. The results are validated through ns-2 simulations. The efficiency of the IDS is analyzed with varying network conditions by simulating routing attacks. The system has achieved an overall detection accuracy of detection with all features is 96.37, rough set feature selection is 97.34 and genetic feature selection is 98.22. Simulation results show that the proposed cross-layer approach aided by a combination of GA and SVM performs significantly better than other approach.
Search Articles
Keywords: Mobile Ad Hoc Networks, Intrusion Detection, Cross-Layer Design, Feature Selection, Rough Set Theory, Genetic Algorithm, Support Vector Machine
Pages: 71-79
WSEAS Transactions on Communications, ISSN / E-ISSN: 1109-2742 / 2224-2864, Volume 13, 2014, Art. #8