2e24c5e8eac125d5b42b21ebd7353b8ec75cc27d,skater/core/global_interpretation/feature_importance.py,FeatureImportance,feature_importance,#FeatureImportance#Any#Any#Any#,19

Before Change

                                       feature_names=self.data_set.feature_ids,
                                       index=self.data_set.index)

        for feature_id in self.data_set.feature_ids:
            // collect perturbations
            if self.data_set.feature_info[feature_id]["numeric"]:
                samples = self.data_set.generate_column_sample(feature_id, n_samples=n, method="stratified")
            else:
                samples = self.data_set.generate_column_sample(feature_id, n_samples=n, method="random-choice")
            copy_of_data_set[feature_id] = samples

            // predict based on perturbed values
            new_predictions = model_instance.predict_subset_classes(copy_of_data_set.data, filter_classes)

            importance = self.compute_importance(new_predictions,
                                                 original_predictions,
                                                 self.data_set[feature_id],
                                                 samples)
            importances[feature_id] = importance

            // reset copy
            copy_of_data_set[feature_id] = self.data_set[feature_id]

        importances = pd.Series(importances).sort_values(ascending=ascending)

        if not importances.sum() > 0:
            self.interpreter.logger.debug("Importances that caused a bug: {}".format(importances))

After Change

        n_jobs = None if n_jobs < 0 else n_jobs
        arg_list = self.data_set.feature_ids
        // just a function of feature_id
        fi_func = partial(input_data=self.data_set.data.copy(),
                          estimator_fn=predict_fn,
                          original_predictions=original_predictions,
                          feature_info=self.data_set.feature_info,
                          feature_names=self.data_set.feature_names,
                          n=n)

        executor_instance = Pool(n_jobs)

        try:

Instances