Unsupervised anomaly detection system using next-generation router architecture
Richard Rouil, Nicolas Chevrollier, Nada Golmie · 2005
Abstract — Unlike many intrusion detection systems that rely mostly on labeled training data, we propose a novel technique for anomaly detection based on unsupervised learning. We apply this technique to counter denial-of-service attacks. Initial simulation results suggest that significant improvements can be obtained. We discuss an implementation of our anomaly detection system in the ForCES router architecture and evaluate it using recorded attack traffic. I.