Search code examples
pythonunsupervised-learninganomaly-detection

Is it possible to explain sklearn isolation forest prediction?

I'm using the isolation forest algorithm from sklearn to do some unsupervised anomaly detection. I need to explained the predictions and I was wondering if there is any way to get the paths that lead to the decision for each sample.

I usually used SHAP or ELI5 but i'd like to do something more custom. So i need the exact path.

Solution

  • You can now use the SHAP values for SkLearn IF : https://github.com/slundberg/shap/pull/784/files