Comparison of different fusion approaches for network intrusion detection using ensemble of RBFNN

• Published in: 2005 International Conference on Machine Learning and Cybernetics Vol. 6; pp. 3846 - 3851 Vol. 6
• Main Authors: Chan, Ng, Yeung, Tsang
• Format: Conference Proceeding
• Language: English
• Published: IEEE 2005
• Subjects: Computer crime; Computer networks; Denial of Service attack; Electronic mail; Fusion; Information technology; Intrusion detection; Machine learning; Multiple Classifier System; Neural Network Ensemble; Neural networks; Radial Basic Function Neural Network; Support vector machine classification; Support vector machines; Voting
• ISBN: 0780390911; 9780780390911
• ISSN: 2160-133X
• DOI: 10.1109/ICMLC.2005.1527610

The information technology has been adopted to solve problems in network intrusion detection system (IDS) and many approaches have been proposed to tackle the information security problems of computer networks, especially the denial of service (DoS) attacks. The multiple classifier system (MCS) is one of the approaches that has been adopted in the detection of DoS attacks recently. For a MCS to perform better than a single classifier, it is crucial to select the appropriate fusion strategies. Majority vote, average, weighted sum, weighted majority vote, neural network and Dempster-Shafer combination are the fusion strategies that have been widely adopted. The selection of the fusion strategy for a MCS in DoS problem varies widely. In this paper, a comparative study on adopting different fusion strategies for a MCS in DoS problem is provided.