[{"data":1,"prerenderedAt":12},["ShallowReactive",2],{"stack-tr-en-shap":3},{"tagline":4,"description":5,"strengths":6,"useCase":11},"The explainability library that attributes every prediction to its features with Shapley values.","SHAP computes, for each prediction, how much every feature contributed, using Shapley values from game theory: the contributions add up exactly to the model's output. TreeExplainer derives them exactly and quickly for XGBoost, LightGBM, CatBoost and random forests; KernelExplainer and DeepExplainer cover any black-box model or neural network. Its plots (waterfall, beeswarm, dependence) make explanations readable by non-specialists, and it's our choice whenever a model has to be audited, justified to business stakeholders or compliant with a regulatory requirement.",[7,8,9,10],"Additive Shapley values: local contributions consistent with the prediction.","TreeExplainer: exact, fast computation for XGBoost, LightGBM, CatBoost, random forests.","KernelExplainer and DeepExplainer for any black-box model or deep network.","Waterfall, beeswarm and dependence plots: from local explanation to global importance.","Explaining credit scores, model audits, regulatory compliance, feature debugging.",1792675979678]