Stacking

polars_ts.ensemble.stacking

Stacking forecaster using a meta-learner.

Trains a meta-learner on out-of-fold predictions from multiple base models, then combines new forecasts through the fitted meta-learner.

StackingForecaster

Stacking ensemble that trains a meta-learner on base model predictions.

Parameters

meta_learner A scikit-learn-compatible estimator with fit and predict. id_col Column identifying each time series. time_col Column with timestamps. target_col Column with actual target values.

fit(cv_predictions, actuals)

Train the meta-learner on out-of-fold base model predictions.

Parameters

cv_predictions List of DataFrames, one per base model, each containing out-of-fold predictions with columns [id_col, time_col, "y_hat"]. actuals DataFrame with actual values in target_col, keyed by id_col and time_col.

Returns

StackingForecaster Fitted forecaster (self).

predict(forecasts)

Combine base model forecasts through the fitted meta-learner.

Parameters

forecasts List of forecast DataFrames, one per base model, each with [id_col, time_col, "y_hat"].

Returns

pl.DataFrame Combined forecast with columns [id_col, time_col, "y_hat"].