StatisticsPerIntervalModel

class StatisticsPerIntervalModel(statistics_function: Callable[[numpy.ndarray], float])

Bases: etna.transforms.decomposition.change_points_based.per_interval_models.base.PerIntervalModel

StatisticsPerIntervalModel gets statistics from series and use them for prediction.

Init StatisticsPerIntervalModel.

Parameters

statistics_function (Callable[[numpy.ndarray], float]) – function to compute statistics from series

Inherited-members

Methods

fit(features, target, *args, **kwargs)	Fit statistics from given target.
predict(features, *args, **kwargs)	Build prediction from precomputed statistics.
set_params(**params)	Return new object instance with modified parameters.
to_dict()	Collect all information about etna object in dict.

fit(features: numpy.ndarray, target: numpy.ndarray, *args, **kwargs) → etna.transforms.decomposition.change_points_based.per_interval_models.statistics_based.StatisticsPerIntervalModel

Fit statistics from given target.

Parameters

• features (numpy.ndarray) – features of the series, will be ignored

• target (numpy.ndarray) – target to compute statistics for

Returns

fitted StatisticsPerIntervalModel

Return type

self

predict(features: numpy.ndarray, *args, **kwargs) → numpy.ndarray

Build prediction from precomputed statistics.

Parameters

features (numpy.ndarray) – features to build prediction for

Returns

array of features len filled with statistics value

Return type

prediction