»Model explainability«
2019-06-03, 14:30–14:55, Hall 2 (PyData)

Model explainability, why should we care. How to explain ML models? SHAP (SHapely Additive exPlanations),

ML-based models are becoming prevalent and affect more and more aspects of our lives. As we rely on models to approve loans, decide whether to hire someone or receive medical treatment, we need to understand why and how the model generates its predictions. What is model explainability? How can we interpret models? SHAP library - How to install - How to generate a global explanation - How to generate a local explanation - Visualizations