baybe.insights.shap.make_explainer_for_surrogate

baybe.insights.shap.make_explainer_for_surrogate(surrogate: Surrogate, data: DataFrame, explainer_cls: type[Explainer] | str = 'KernelExplainer', *, use_comp_rep: bool = False)[source]

Create a SHAP explainer for a given surrogate model.

Parameters:

  • surrogate (Surrogate) – The surrogate model to be explained.

  • data (DataFrame) – The background data set.

  • explainer_cls (type[Explainer] | str) – The SHAP explainer class for generating the explanation.

  • use_comp_rep (bool) – Boolean flag specifying whether to explain the model’s experimental or computational representation.

Return type:

Explainer

Returns:

The created explainer object.

Raises:

  • ValueError – If the provided background data set is empty.

  • TypeError – If the provided explainer class is incompatible with the surrogate.