StackingForecaster

StackingForecaster.

Quickstart

from sktime.forecasting.compose import StackingForecaster

estimator = StackingForecaster(forecasters, regressor=None, random_state=None, n_jobs=None)

Parameters(4)

forecasterslist of (str, estimator) tuples

Estimators to apply to the input series.

regressor: sklearn-like regressor, optional, default=None.

The regressor is used as a meta-model and trained with the predictions of the ensemble forecasters as exog data and with y as endog data. The length of the data is dependent to the given fh. If None, then a GradientBoostingRegressor(max_depth=5) is used. The regressor can also be a sklearn.Pipeline().

random_stateint, RandomState instance or None, default=None

Used to set random_state of the default regressor.

n_jobsint or None, optional (default=None)

The number of jobs to run in parallel for fit. None means 1 unless in a joblib.parallel_backend context. -1 means using all processors.

Examples

>>> from sktime.forecasting.compose import StackingForecaster
>>> from sktime.forecasting.naive import NaiveForecaster
>>> from sktime.forecasting.trend import PolynomialTrendForecaster
>>> from sktime.datasets import load_airline
>>> y = load_airline ()
>>> forecasters = [
... ("trend", PolynomialTrendForecaster ()),
... ("naive", NaiveForecaster ()),
... ]
>>> forecaster = StackingForecaster (forecasters = forecasters)
>>> forecaster. fit (y = y, fh = [1, 2, 3 ]) StackingForecaster(
... )
>>> y_pred = forecaster. predict ()