tsad.utils.ResidualAnomalyDetectionUtils package

Submodules

tsad.utils.ResidualAnomalyDetectionUtils.feature_importance module

tsad.utils.ResidualAnomalyDetectionUtils.feature_importance.feature_importance(residuals, analysis_type='collective', date_from=None, date_till=None, weigh=True)

Feature importance calculation

Parameters:

residualspandas.DataFrame()

analysis_typestr, “single”/”collective”, “single” by default

date_fromstr in format ‘yyyy-mm-dd HH:MM:SS’, None by default

date_tillstr in format ‘yyyy-mm-dd HH:MM:SS’, None by default

weighboolean, True by default

If analysis_type == “collective”.

Returns:

datapandas.DataFrame().

tsad.utils.ResidualAnomalyDetectionUtils.generateResidual module

tsad.utils.ResidualAnomalyDetectionUtils.generateResidual.absoluteResidual(y_pred, y_true)

Функция позволяющая получить разницу

tsad.utils.ResidualAnomalyDetectionUtils.stastics module

class tsad.utils.ResidualAnomalyDetectionUtils.stastics.Hotelling(koef_ucl=3)

Bases: object

Methods

feature_importances
fit
fit_predict
predict

feature_importances(df)

fit(df)

fit_predict(df, show_figure=False)

predict(df, show_figure=False)

Module contents