energy_fault_detector.root_cause_analysis.arcana_utils

calculate_arcana_importance_time_series(bias_data)

Calculate ARCANA importances for each time stamp.

Parameters:

bias_data (DataFrame) – pandas DataFrame containing ARCANA bias data, with timestamps as index.

Return type:

DataFrame

Returns:

pandas DataFrame with time stamps as index and importances for each feature as values.

calculate_mean_arcana_importances(bias_data, start=None, end=None)

Calculate the mean ARCANA importances for a given period. If normal_index is provided, only the timestamps during normal status are considered.

ARCANA importances express the contribution of each data feature to the resulting reconstruction error by comparing the provided bias data.

Parameters:

• bias_data (DataFrame) – pandas DataFrame containing ARCANA bias data, with timestamps as index.

• start (Union[str, datetime]) – start of time period to evaluate.

• end (Union[str, datetime]) – end of time period to evaluate.

Return type:

Series

Returns:

pandas Series with features/column names as index and importances as values.