Forecaster
DummyGlobalForecaster
Dummy global forecaster that predicts mean of pretrain data.
This forecaster implements pretraining by computing the mean of all time series in the pretrain set, then predicts that mean for all future time points.
Useful as:
Simple baseline for global forecasting
Test case for the pretraining API
Educational examples
Schnellstart
python
from sktime.forecasting.dummy_global import DummyGlobalForecaster
estimator = DummyGlobalForecaster(strategy='mean')Parameter(1)
- strategystr, one of {“mean”, “last”, “mean_by_index”}, default=”mean”
Strategy for prediction:
"mean": predict mean of all values in pretrain set"last": predict last value from fit data"mean_by_index": predict mean computed per time index across pretraining series. Useful for cold start scenarios where pattern by index matters.
Beispiele
>>> from sktime.forecasting.dummy_global import DummyGlobalForecaster
>>> from sktime.utils._testing.hierarchical import _make_hierarchical
>>> # Create panel of training data
>>> y_panel = _make_hierarchical (
... hierarchy_levels = (2,), min_timepoints = 10, max_timepoints = 10
... )
>>> forecaster = DummyGlobalForecaster ()
>>> forecaster. pretrain (y_panel) # Learn global mean DummyGlobalForecaster()
>>> # Now fit to a specific series
>>> from sktime.datasets import load_airline
>>> y = load_airline ()
>>> forecaster. fit (y) # Set context DummyGlobalForecaster()
>>> y_pred = forecaster. predict (fh = [1, 2, 3 ]) # Predict global mean
>>> y_pred. shape (3,)