OCDocker.OCScore.Utils.Evaluation module

Set of functions to manage evaluation operations in OCDocker in the context of scoring functions.

Usage:

import OCDocker.OCScore.Utils.Evaluation as ocseval

OCDocker.OCScore.Utils.Evaluation.compute_auc(df, positive_class_names, score_columns, class_column_name)[source]

Compute the AUC for the scores in given score_columns.

Parameters:

  • df (pd.DataFrame) – The DataFrame containing the scores.

  • positive_class_names (Union[str, list[str]]) – The name/names of the positive class. If a string is given, it will be converted to a list. (All other classes will be considered as negative)

  • score_columns (list[str]) – The list of columns containing the scores.

  • class_column_name (str) – The name of the column containing the class.

Returns:

The DataFrame with the computed AUC.

Return type:

pd.DataFrame

OCDocker.OCScore.Utils.Evaluation.compute_metrics(df, score_columns, target_column_name, db_column_name, metric_db_name, class_column_name, positive_class_names, invert_conditionally=True)[source]

Compute the metrics for the scores in given score_columns.

Parameters:

  • df (pd.DataFrame) – The DataFrame containing the scores.

  • score_columns (list[str]) – The list of columns containing the scores.

  • target_column_name (str) – The name of the column containing the target values.

  • db_column_name (str) – The name of the column containing the database name which will be assigned to RMSE or AUC.

  • metric_db_name (tuple[str, str]) – The pair of values which will be assigned to RMSE and AUC respectively.

  • class_column_name (str) – The name of the column containing the class.

  • positive_class_names (Union[str, list[str]]) – The name/names of the positive class. If a string is given, it will be converted to a list. (All other classes will be considered as negative)

  • invert_conditionally (bool) –

Returns:

The DataFrame with the computed metrics.

Return type:

pd.DataFrame

Raises:

ValueError – If metric_db_name does not have two elements.

OCDocker.OCScore.Utils.Evaluation.compute_rmse(df, score_columns, target_column_name)[source]

Compute the RMSE for the scores in given score_columns.

Parameters:

  • df (pd.DataFrame) – The DataFrame containing the scores.

  • score_columns (list[str]) – The list of columns containing the scores.

  • target_column_name (str) – The name of the column containing the target values.

Returns:

The DataFrame with the computed RMSE.

Return type:

pd.DataFrame