get_median_shap_contribs#

openstef_models.utils.xgboost.get_median_shap_contribs(booster: Booster, input_data: DataFrame, quantiles: list[Quantile]) → ndarray[source]#

Compute SHAP contributions and return the median-quantile slice.

Parameters:

  • booster (Booster) – Fitted XGBoost Booster.

  • input_data (DataFrame) – Feature matrix for which to compute contributions.

  • quantiles (list[Quantile]) – Quantiles the model was trained on.

  • booster

  • input_data

  • quantiles

Returns:

Array of shape (n_samples, n_features + 1) where the last column is the model bias.

Return type:

ndarray