Conditional expectation network for SHAP
This work addresses a computational bottleneck for researchers and practitioners using SHAP for model interpretability, offering an incremental improvement over existing methods.
The authors tackled the computational inefficiency of conditional SHAP explanations by proposing a surrogate neural network approach that efficiently calculates conditional SHAP for neural networks and other regression models, while properly accounting for feature dependencies. This method also enables drop1, ANOVA analyses, and improved partial dependence plots in complex models.
A very popular model-agnostic technique for explaining predictive models is the SHapley Additive exPlanation (SHAP). The two most popular versions of SHAP are a conditional expectation version and an unconditional expectation version (the latter is also known as interventional SHAP). Except for tree-based methods, usually the unconditional version is used (for computational reasons). We provide a (surrogate) neural network approach which allows us to efficiently calculate the conditional version for both neural networks and other regression models, and which properly considers the dependence structure in the feature components. This proposal is also useful to provide drop1 and anova analyses in complex regression models which are similar to their generalized linear model (GLM) counterparts, and we provide a partial dependence plot (PDP) counterpart that considers the right dependence structure in the feature components.