Repository navigation
[ENH] skforecast integration #199
Description
It would be nice to have an estimator level integration with detectors in the skforecast package.
https://github.com/skforecast/skforecast
This would require:
- a
BaseExperimentdescendant classSkforecastExperimentwhich carries out a backtesting experiment for time series forecasters. This can lean onSktimeForecastingExperiment. Also see the extension template https://github.com/SimonBlanke/Hyperactive/blob/main/extension_templates/experiments.py - a
SkforecastOptCvclass inheriting fromskforecastBaseEstimatorsimilar to the other classes inhyperactive.integrations. Internally, it uses theSkforecastExperiment
Activity
SimonBlanke commentedon Sep 27, 2025
Same questions as the "skpro"-issue:
Very interesting! Could you tell me about possible applications. Do you already have packages or projects in mind where this integration will be used (GC.OS software stack)?
fkiraly commentedon Sep 27, 2025
I thought simple tuning on the level of
skforecastestimators might be nice. @JoaquinAmatRodrigo and @JavierEscobarOrtiz might be able to say whether they would like such functionality.FYI, I think with
hyperactiveit is now extremely easy to write tuners for any AI framework! Assuming there already is some basic benchmarking functionality that can be wrapped by aBaseExperiment.staru09 commentedon Nov 15, 2025
Would love to work on this if it's available.
fkiraly commentedon Nov 16, 2025
@staru09, yes, it is available!
staru09 commentedon Nov 17, 2025
Please assign this to me.