Glossary · Business Analytics

Isolation Forest

also: iForest · tree-based anomaly detection

Definition

Isolation Forest is a tree-based anomaly detection algorithm that scores observations by how easily they can be isolated via random recursive partitioning. Anomalies are isolated in few splits; normal points require many. The algorithm handles mixed feature types without density estimation or distance calculations.

Liu, Ting and Zhou (2008) introduced isolation forest as an alternative to density-based outlier detection. Trees are built by random feature and split selection; path length (number of splits to isolate a point) is the anomaly score. The method scales linearly in sample size, handles high-dimensional data, and does not require anomaly examples to train. In revenue anomaly detection it complements Prophet-style decomposition by flagging pattern-breaking observations the time-series model cannot explain.

Essays on this concept