From Log-Odds to Shapley Values: An Explanatory Geometry for the Weighted Naive Bayes Classifier

This paper studies the construction of an explanatory space for a weighted naive Bayes classifier from the supervised representation induced by the model.