arXivML & AI Methods

Case-based Explainability for Random Forest: Prototypes, Critics, Counter-factuals and Semi-factuals

The study explores Explainable Case-Based Reasoning methods to clarify the results of black-box machine learning algorithms, using a technique to extract the learned distance metric from Random Forests, and assesses their explanatory power.