History

skpro was started in 2017 by Franz Király and his then-student Frithjof Gressmann as a scikit-learn like python package for probabilistic supervised regression.

skpro was then abandoned, from 2019, at version 1.0.1, as development in Franz Király research group continued to be focused on sktime.

In 2022-23, sktime’s base module was turned into a separate package, skbase, intended as a workbench to allow easy templating and creation of scikit-learn-likes.

Using the templating scaffold of skbase, skpro was finally revived in 2023 by Franz Király, Frithjof Gressmann, Anirban Ray, and Alex Gregory, built upon a fully rearchitectured, skbase reliant API, as version 2.0.0.

The joint base interface enables mutual compabitibilty between skpro, sklearn, and sktime, with skpro probabilistic regressors being potential components used for probabilistic forecasting in sktime.

Development is supported by members of the sktime project, new core developers and the broader community (see contributors).

If you are interested in contributing, check out our Contributing guide.