Proceedings of the 11th International Symposium on Recent Advances in Intrusion Detection— Posters
Information-theoretic metrics hold great promise for modeling traffic and detecting anomalies if only they could be computed in an efficient, scalable ways. Recent advances in streaming estimation algorithms give hope that such computations can be made practical. We describe our work in progress that aims to use streaming algorithms on 802.11a/b/g link layer (and above) features and feature pairs to detect anomalies.
Sergey Bratus, Joshua Brody, David Kotz, and Anna Shubina. Streaming Estimation of Information-theoretic Metrics for Anomaly Detection (Extended Abstract). In Proceedings of the 11th International Symposium on Recent Advances in Intrusion Detection— Posters, September 2008. 10.1007/978-3-540-87403-4_32