get_ttest_summary_df

get_ttest_summary_df#

mlproject.postprocess.utils.get_ttest_summary_df(target_name, models_dir, num_folds=10, model_type='rf', feature_set_types=['matminer', 'matminer_lob'], alternative='two-sided')[source]#

Get t-test model summary dataframe including effect size and relative improvement.

Parameters:

  • target_name (str) – Target property name.

  • models_dir (str) – Base directory containing model results.

  • num_folds (int, optional) – Number of CV folds. Default is 10.

  • model_type (str, optional) – Model name/prefix (e.g., ‘rf’, ‘modnet’). Default is ‘rf’.

  • feature_set_types (list of str, optional) – List of feature set variants to compare. Default is [‘matminer’, ‘matminer_lob’]

  • alternative (str)

Returns:

summary_df – Summary dataframe with t-test results, effect size, and relative improvement.

Return type:

pd.DataFrame

Contents