pyanno4rt.learning_model.inspection.algorithms

Inspection algorithms module.

The module aims to provide functions to run different inspection algorithms.

Overview

Function

permutation_importances(model_instance, hyperparameters, features, labels, preprocessing_steps, number_of_repeats, oof_folds)

Compute the permutation importances.

Functions

pyanno4rt.learning_model.inspection.algorithms.permutation_importances(model_instance, hyperparameters, features, labels, preprocessing_steps, number_of_repeats, oof_folds)[source]

Compute the permutation importances.

Parameters:

  • model_instance (object) – The object representing the machine learning outcome model.

  • hyperparameters (dict) – Dictionary with the machine learning outcome model hyperparameters.

  • features (ndarray) – Values of the input features.

  • labels (ndarray) – Values of the input labels.

  • preprocessing_steps (list) – Sequence of labels associated with the preprocessing algorithms for the machine learning outcome model.

  • number_of_repeats (int) – Number of feature permutations.

  • oof_folds (ndarray) – Out-of-fold split numbers.

Returns:

Dictionary with the training and out-of-folds permutation importances.

Return type:

dict