postprocessing

Forecast postprocessing transformations.

Contains transforms that are applied to forecast results to improve accuracy, apply business constraints, or enhance prediction quality. These transforms operate on ForecastDataset objects after the core prediction step.

Submodules

openstef_models.transforms.postprocessing.confidence_interval_applicator	Confidence interval generation for probabilistic forecasts.
openstef_models.transforms.postprocessing.conformalized_quantile_calibrator	Asymmetric conformal quantile calibration without external dependencies.
openstef_models.transforms.postprocessing.isotonic_quantile_calibrator	Isotonic quantile calibration for probabilistic forecasts.
openstef_models.transforms.postprocessing.quantile_sorter	Quantile forecast ordering correction.

Classes

ConfidenceIntervalApplicator(**data)	Add quantile predictions to forecasts based on learned uncertainty patterns.
ConformalizedQuantileCalibrator(**data)	Apply asymmetric split-conformal corrections to forecast quantiles.
IsotonicQuantileCalibrator(**data)	Calibrate quantile predictions using isotonic regression.
QuantileSorter()	Sort quantile forecasts to enforce monotonic ordering.