59. Explainable machine learning
Explain individual predictions and global model behavior with coefficients, feature importance, partial dependence, local attribution, examples, and counterfactuals. Match each explanation method to its assumptions, audience, and failure modes.