Stuck on an issue?

Lightrun Answers was designed to reduce the constant googling that comes with debugging 3rd party libraries. It collects links to all the places you might be looking at while hunting down a tough bug.

And, if you’re still stuck at the end, we’re happy to hop on a call to see how we can help out.

Forecasters to have a default ForecastingHorizon of fh=[1] if fh is not given (None)

See original GitHub issue

Is your feature request related to a problem? Please describe. I would like to propose to have a default 1-step ahead fh for all forecasters. Motivation is, that 1-step ahead forecast is often used and a default would simplify the interface and UX.

Describe the solution you’d like We could just catch cases where fh is None and set it to fh=[1]. We have to see how this works with absolute fh.

forecaster = NaiveForecaster()
forecaster.fit(y)
forecaster.predict()

Martin Walter martin_friedrich.walter@daimler.com, Mercedes-Benz AG on behalf of Daimler TSS GmbH. https://github.com/Daimler/daimler-foss/blob/master/LEGAL_IMPRINT.md

Issue Analytics

State:
Created 2 years ago
Comments:19 (2 by maintainers)

Top GitHub Comments

2reactions

TonyBagnallcommented, Jul 5, 2021

I’m just butting in, feel free to ignore me, but this seems to have gone from being a simple enhancement to something quite complex. Perhaps drop the idea for now?

1reaction

aiwaltercommented, Jul 9, 2021

agree, seems the impact of implementing this is much bigger than the benefit it will bring. So I will close this issue for now. Thx anyway for the discussion.

Top Results From Across the Web

sktime/_base.py at main - GitHub

fh : int, list, np.array or ForecastingHorizon, optional (default=None). The forecasting horizon encoding the time stamps to forecast at. if has not been ......

ForecastingHorizon — sktime documentation

Whether the forecasting horizon is purely out-of-sample for given cutoff. Parameters. cutoffpd.Period, pd.Timestamp, int, optional (default=None). Cutoff value ...

Why start using sktime for forecasting? - Towards Data Science

The relative forecasting horizon is especially useful if we make rolling predictions using the update_predict() method.

sktime.utils.validation.check_window_length Example

Series Target time series to which to fit the forecaster. X : pd.DataFrame, optional (default=None) Exogenous variables are ignored fh : int, ...

Cross-validation results for the selected models with the ...

Download scientific diagram | Cross-validation results for the selected models with the forecasting horizon (FH) of 1-4 quarters and corresponding ...