|
|
Isolation Forest for anomaly detection |
|
|
Most existing model-based approaches to anomaly detection construct a profile of normal instances, then identify instances that do not conform to the normal profile as anomalies. This talk is about a fundamentally different model-based method that explicitly isolates anomalies instead of profiles normal points. To our best knowledge, the concept of isolation has not been explored in current literature. The use of isolation enables the proposed method, iForest, to exploit sub-sampling to an extent that is not feasible in existing methods, creating an algorithm which has a linear time complexity with a low constant and a low memory requirement. Our empirical evaluation shows that iForest performs favourably to ORCA, a near-linear time complexity distance-based method, LOF and Random Forests in terms of AUC and processing time, and especially in large data sets. |
|
|
Tony is married to Jackie and has three children; they are Enoch, Philip and Anna. Tony has over 10 years experience in commerical software development and he is now in the final year of his PhD candidature. In 2006, Tony is awarded with Best Paper (and Best Student Paper) in PAKDD and in 2008, Runner-up Best Theoretical/Algorithms Paper Award in ICDM. Tony also has reviewer duty in various journal and conference publications, for examples, Journal of machines learning Research and, Knowledge and Information Systems: An International Journal. |
|
