SHapley Additive exPlanations
SHapley Additive exPlanations (SHAP) is a framework to interpret the results of any machine learning or statistical model[1][2]. It is based on the game theoretical concept of Shapley value.
The basic idea of SHAP is to first approximate the model with a locally accurate (i.e. accurate for a single input) linear explanation model with binary features and then compute the Shapley value for each of the features. In general, this task has been shown to be NP-hard[3]. However, several algorithms have been developed to approximate ...
References
- ↑ Lundberg, Scott M; Lee, Su-In (2017). "A unified approach to interpreting model predictions". Advances in Neural Information Processing Systems. 30.
- ↑ Ponce-Bobadilla, Ana Victoria; Schmitt, Vanessa; Maier, Corinna S; Mensing, Sven; Stodtmann, Sven (2024). "Practical guide to SHAP analysis: Explaining supervised machine learning model predictions in drug development". Clinical and Translational Science. Wiley Online Library. 17 (11): e70056. doi:10.1111/cts.70056. PMC 11513550 Check
|pmc=value (help). PMID 39463176 Check|pmid=value (help). - ↑ Matsui, Yasuko; Matsui, Tomomi (2001). "NP-completeness for calculating power indices of weighted majority games". Theoretical Computer Science. Elsevier. 263 (1–2): 305–310. doi:10.1016/S0304-3975(00)00261-7 (inactive 15 October 2025).
This article "SHapley Additive exPlanations" is from Wikipedia. The list of its authors can be seen in its historical and/or the page Edithistory:SHapley Additive exPlanations. Articles copied from Draft Namespace on Wikipedia could be seen on the Draft Namespace of Wikipedia and not main one.
