Statistical Aspects of SHAP: Functional ANOVA for Model Interpretation. (arXiv:2208.09970v1 [stat.ME])

SHAP is a popular method for measuring variable importance in machine
learning models. In this paper, we study the algorithm used to estimate SHAP
scores and show that it is a transformation of the functional ANOVA
decomposition. We use this connection to show that challenges in SHAP
approximations largely relate to the choice of a feature distribution and the
number of $2^p$ ANOVA terms estimated. We argue that the connection between
machine learning explainability and sensitivity analysis is illuminating in …

anova arxiv interpretation shap statistical