This is the supporting notebook for the research paper — Why Are You Weird? Infusing Interpretability in Isolation Forest for Anomaly Detection — published in the Explainable AI Workshop Proceedings of the 35th AAAI Conference, 2021. Pre-print version is available on arxiv.

Both commented code of the experiments and the results are reproduced at full in the AWS_demo notebook.

Since this is the first release of the implementation of the Assist-Based Weighting Scheme (AWS), which powers our model-specific local explanation method for the Isolation Forest learner, we welcome all constructive feedback and suggestions.