pdmlabs.method.semi_supervised_method

pdmlabs.method.semi_supervised_method#

Classes

SemiSupervisedMethodInterface(event_preferences)

class pdmlabs.method.semi_supervised_method.SemiSupervisedMethodInterface(event_preferences: EventPreferences)#

Bases: MethodInterface

abstract fit(historic_data: list[DataFrame], historic_sources: list[str], event_data: DataFrame) → None#

This method is used to fit a anomaly detection model in (relative) normal data, where the data are passed in form of Dataframes along with their respected source.

Parameters:

  • historic_data – a list of Dataframes (used to fit a semi-supervised model). The historic_data list parameter elements should be copied if a corresponding method needs to store them for future processing

  • historic_sources – a list with strings (names) of the different sources

  • event_data – event data that are produced from the different sources

Returns:

None.

Contents