Skip to content

Feature Importance and Explanations

MPSBoost provides multiple explanation layers:

  • gain importance
  • split-count importance
  • permutation importance
  • controlled SHAP-like approximate explanations

Design principles

  • Do not duplicate prediction or scoring logic.
  • Do not present approximate explanations as official SHAP.
  • Advance the official SHAP TreeExplainer adapter as a separate task.

Usage direction

When research or reporting needs official SHAP semantics, wait for the S16.5a / S23 adapter documentation. For the current package, use permutation_importance as model-agnostic interpretation.

Official SHAP path

Official SHAP is optional and explicit:

python -m pip install 'mpsboost[shap]'

export_native_trees_for_shap(estimator) exports native tree structure for TreeExplainer adapter validation without training data, credentials, telemetry, or device identifiers. official_shap_tree_explainer(...) stops clearly until semantic validation is enabled, so approximate explanations are not presented as official SHAP output.