Conformal Prediction in Python with crepes

Henrik Boström
Proceedings of the Thirteenth Symposium on Conformal and Probabilistic Prediction with Applications, PMLR 230:236-249, 2024.

Abstract

\verb|crepes| is a Python package for conformal prediction, which has been extended in several ways since its introduction. While the original version of the package focused on conformal regressors and predictive systems, the current version also includes conformal classifiers. New classes and methods for computing non-conformity scores and Mondrian categories have also been incorporated. Moreover, the package has been extended to allow for seamless embedding of classifiers and regressors in the conformal prediction framework; instead of generating conformal predictors that are separate from the learners, the latter can now be equipped with specific prediction methods that in addition to providing point predictions also can generate p-values, prediction sets and intervals, as well as conformal predictive distributions. Extensive documentation for the package has furthermore been developed. In this paper, these extensions are described, as implemented in \verb|crepes|, version 0.7.0.

Cite this Paper

BibTeX
@InProceedings{pmlr-v230-bostrom24a, title = {Conformal Prediction in Python with crepes}, author = {Bostr\"{o}m, Henrik}, booktitle = {Proceedings of the Thirteenth Symposium on Conformal and Probabilistic Prediction with Applications}, pages = {236--249}, year = {2024}, editor = {Vantini, Simone and Fontana, Matteo and Solari, Aldo and Boström, Henrik and Carlsson, Lars}, volume = {230}, series = {Proceedings of Machine Learning Research}, month = {09--11 Sep}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v230/main/assets/bostrom24a/bostrom24a.pdf}, url = {https://proceedings.mlr.press/v230/bostrom24a.html}, abstract = {\verb|crepes| is a Python package for conformal prediction, which has been extended in several ways since its introduction. While the original version of the package focused on conformal regressors and predictive systems, the current version also includes conformal classifiers. New classes and methods for computing non-conformity scores and Mondrian categories have also been incorporated. Moreover, the package has been extended to allow for seamless embedding of classifiers and regressors in the conformal prediction framework; instead of generating conformal predictors that are separate from the learners, the latter can now be equipped with specific prediction methods that in addition to providing point predictions also can generate p-values, prediction sets and intervals, as well as conformal predictive distributions. Extensive documentation for the package has furthermore been developed. In this paper, these extensions are described, as implemented in \verb|crepes|, version 0.7.0.} }
Endnote
%0 Conference Paper %T Conformal Prediction in Python with crepes %A Henrik Boström %B Proceedings of the Thirteenth Symposium on Conformal and Probabilistic Prediction with Applications %C Proceedings of Machine Learning Research %D 2024 %E Simone Vantini %E Matteo Fontana %E Aldo Solari %E Henrik Boström %E Lars Carlsson %F pmlr-v230-bostrom24a %I PMLR %P 236--249 %U https://proceedings.mlr.press/v230/bostrom24a.html %V 230 %X \verb|crepes| is a Python package for conformal prediction, which has been extended in several ways since its introduction. While the original version of the package focused on conformal regressors and predictive systems, the current version also includes conformal classifiers. New classes and methods for computing non-conformity scores and Mondrian categories have also been incorporated. Moreover, the package has been extended to allow for seamless embedding of classifiers and regressors in the conformal prediction framework; instead of generating conformal predictors that are separate from the learners, the latter can now be equipped with specific prediction methods that in addition to providing point predictions also can generate p-values, prediction sets and intervals, as well as conformal predictive distributions. Extensive documentation for the package has furthermore been developed. In this paper, these extensions are described, as implemented in \verb|crepes|, version 0.7.0.
APA
Boström, H.. (2024). Conformal Prediction in Python with crepes. Proceedings of the Thirteenth Symposium on Conformal and Probabilistic Prediction with Applications, in Proceedings of Machine Learning Research 230:236-249 Available from https://proceedings.mlr.press/v230/bostrom24a.html.

Related Material