kats.detectors.bocpd_model module¶

This file implements the Bayesian Online Changepoint Detection algorithm as a DetectorModel, to provide a common interface.

class kats.detectors.bocpd_model.BocpdDetectorModel(serialized_model: Optional[bytes] = None, slow_drift: bool = False, threshold: Optional[float] = None)[source]¶

Bases: kats.detectors.detector.DetectorModel

Implements the Bayesian Online Changepoint Detection as a DetectorModel.

This provides an unified interface, which is common to all detection algorithms.

serialized_model¶

json containing information about stored model.

slow_drift¶

Boolean. True indicates we are trying to detect trend changes. False indicates we are trying to detect level changes.

Typical Usage: level_ts is an instance of TimeSeriesData >>> bocpd_detector = BocpdDetectorModel() >>> anom = bocpd_detector.fit_predict(data=level_ts)

fit(data: kats.consts.TimeSeriesData, historical_data: Optional[kats.consts.TimeSeriesData] = None) → None[source]¶

fit can be called during priming. It’s a noop for us.

fit_predict(data: kats.consts.TimeSeriesData, historical_data: Optional[kats.consts.TimeSeriesData] = None) → kats.detectors.detector_consts.AnomalyResponse[source]¶

Finds changepoints and returns score.

Uses the current data and historical data to find the changepoints, and returns an AnomalyResponse object, the scores corresponding to probability of changepoints.

Parameters

  • data – TimeSeriesData object representing the data

  • historical_data – TimeSeriesdata object representing the history. Dats

  • start exactly where the historical_data ends. (should) –

Returns

AnomalyResponse object, representing the changepoint probabilities. The score property contains the changepoint probabilities. The length of the object is the same as the length of the data.

predict(data: kats.consts.TimeSeriesData, historical_data: Optional[kats.consts.TimeSeriesData] = None) → kats.detectors.detector_consts.AnomalyResponse[source]¶

predict is not implemented

serialize() → bytes[source]¶

Returns the serialzed model.

Parameters

None. –

Returns

json containing information about serialized model.