Is Shapley Explanation for a model unique?
This work highlights a critical limitation in interpretability for practitioners using Shapley values, showing that explanations are not unique and depend on model application, which is incremental but important for reliability.
The paper investigates the uniqueness of Shapley value explanations for machine learning models, finding that explanations vary based on the model's predicted outcomes (e.g., probability, log-odds, binary decisions) and feature distribution moments like variance, leading to disagreements in feature importance.
Shapley value has recently become a popular way to explain the predictions of complex and simple machine learning models. This paper is discusses the factors that influence Shapley value. In particular, we explore the relationship between the distribution of a feature and its Shapley value. We extend our analysis by discussing the difference that arises in Shapley explanation for different predicted outcomes from the same model. Our assessment is that Shapley value for particular feature not only depends on its expected mean but on other moments as well such as variance and there are disagreements for baseline prediction, disagreements for signs and most important feature for different outcomes such as probability, log odds, and binary decision generated using same linear probability model (logit/probit). These disagreements not only stay for local explainability but also affect the global feature importance. We conclude that there is no unique Shapley explanation for a given model. It varies with model outcome (Probability/Log-odds/binary decision such as accept vs reject) and hence model application.