src.eval.rand_forestpp_evaluator

Classes

RandForestPPEvaluator(**kwargs)

class src.eval.rand_forestpp_evaluator.RandForestPPEvaluator(**kwargs)
Author:

Alberto M. Esmoris Pena

Class to evaluate trained C++ random forest models. See RandomForestPPClassificationModel.

The evaluator computes MDI feature importances from the C++ backend and optionally computes permutation importance. Unlike the sklearn evaluator, tree plotting produces feature importance bar charts instead of tree diagrams.

Variables:

  • compute_permutation_importance (bool) – Whether to compute permutation importance (True) or not (False).

  • n_repeats (int) – Number of shuffle repeats for permutation importance.

__init__(**kwargs)

Initialize/instantiate a RandForestPPEvaluator.

Parameters:

kwargs – The attributes for the RandForestPPEvaluator.

eval(model, X=None, y=None, **kwargs)

Evaluate a trained C++ random forest model.

Parameters:

  • model (RandomForestPPClassificationModel) – The C++ random forest model.

  • X – The matrix of point-wise features. Rows are points, columns are features.

  • y – The vector of classes.

Returns:

The evaluation of the trained C++ random forest.

Return type:

RandForestPPEvaluation