SHAP for additively modeled features in a boosted trees model
This provides a theoretical insight for interpretability in machine learning, but it is incremental as it builds on existing SHAP and partial dependence methods.
The paper tackled the relationship between SHAP dependence plots and partial dependence plots for additively modeled features in boosted trees models, showing they correspond up to a vertical shift, as illustrated with XGBoost.
An important technique to explore a black-box machine learning (ML) model is called SHAP (SHapley Additive exPlanation). SHAP values decompose predictions into contributions of the features in a fair way. We will show that for a boosted trees model with some or all features being additively modeled, the SHAP dependence plot of such a feature corresponds to its partial dependence plot up to a vertical shift. We illustrate the result with XGBoost.