The Nomination for Reducing IDS false positives using Incremental Stream Clustering (ISC) Algorithm by Champa Dey

Champa Dey has during her master thesis work at SICS managed to read in on a large number of subjects related to AI, like machine learning, Bayesian statistics, anomaly detection, and intrusion detection. She has used that knowledge to further develop the Incremental Stream Clustering methods for statistical anomaly detection originally developed at SICS, and applied it on the area of alarm filtering for intrusion detection in computer networks, with very good results.

Intrusion Detection is an ever more important area, where AI methods can be very useful, and with high potential for making it into real applications. Champa Dey has indeed shown that statistical anomaly detection is a promising approach in this area.

Institution: DSV/KTH and SICS
Supervisors: Louise Yngström (DSV); Tomas Olsson and Anders Holst (SICS)

Anders Holst, SICS