Conformalized selective regression
Discover Data · 29.07.2026 · 4 (1)
DOI: 10.1007/s44248-026-00113-2 ↗Kurz gefasst
Fachartikel in der Zeitschrift Discover Data, online erschienen am 29.07.2026. Verfasst von Anna Sokol und 2 weiteren Autor:innen. Die Arbeit nennt 30 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY-NC-ND 4.0).
Abstract (Originalsprache)
Should prediction models always deliver a prediction? In the pursuit of maximum predictive performance, critical considerations of reliability are often overshadowed, particularly when it comes to the role of uncertainty. Selective regression, also known as the “reject option,” allows models to abstain from predictions in cases of considerable uncertainty. Initially proposed seven decades ago, approaches to selective regression have mostly focused on distribution-based proxies for measuring uncertainty, particularly conditional variance. However, this focus neglects the significant influence of model-specific biases on performance. In this paper, we propose a novel approach to selective regression by leveraging conformal prediction, which provides grounded confidence measures for individual predictions based on model-specific biases. In addition, we propose a standardized evaluation framework to allow proper comparison of selective regression approaches. Via an extensive experimental approach, we demonstrate how our proposed approach, conformalized selective regression, presents an advantage over multiple state-of-the-art comparison models.
Zitieren
APA
Sokol, A., Moniz, N., & Chawla, N. V. (2026). Conformalized selective regression. Discover Data, 4(1). https://doi.org/10.1007/s44248-026-00113-2
Vancouver
Sokol A, Moniz N, Chawla NV. Conformalized selective regression. Discover Data. 2026;4(1). doi:10.1007/s44248-026-00113-2
BibTeX
@article{Sokol2026,
author = {Sokol, Anna and Moniz, Nuno and Chawla, Nitesh V.},
title = {Conformalized selective regression},
journal = {Discover Data},
year = {2026},
volume = {4},
number = {1},
doi = {10.1007/s44248-026-00113-2}
}
EUVIMED