calculate_permutation_feature_importance {GeneSelectR}

Calculate Permutation Feature Importance

Description

This function calculates permutation feature importance for a Scikit-learn pipeline with a trained classifier as the final step.

Usage

calculate_permutation_feature_importance(
  pipeline,
  X_train,
  y_train,
  n_repeats = 10L,
  random_state = 0L,
  njobs = njobs,
  pipeline_name,
  iter
)

Arguments

pipeline

A Scikit-learn pipeline object with a trained classifier as the final step.

X_train

A DataFrame containing the training data.

y_train

A DataFrame containing the training labels.

n_repeats

An integer specifying the number of times to permute each feature.

random_state

An integer specifying the seed for the random number generator.

njobs

An integer specifying number of cores to use. Set up by the master GeneSelectR function.

pipeline_name

Strings (names of the selected_pipelines list) representing pipeline names that were constructed for the feature selection

iter

An integer that is indicating current iteration of the train-test split

Value

A dataframe containing the feature names and their permutation importance scores, ranked by importance. Each row represents a feature, with columns for feature names, importances, and ranks.