Inductive Venn–Abers and related regressors

Vladimir Vovk; Ivan Petej

Inductive Venn–Abers and related regressors

Vladimir Vovk
, Ivan Petej

Department of Computer Science

Research output: Chapter in Book/Report/Conference proceeding › Chapter (peer-reviewed) › peer-review

Abstract

Venn–Abers predictors are probabilistic predictors that enjoy appealing properties of validity, but their major limitation is that they are applicable only to the case of binary classification, with a recent extension to bounded regression. We generalize them to the case of unbounded regression, which requires adding an element of conformal prediction. In our simulation and empirical studies we investigate the predictive efficiency of point regressors derived from Venn–Abers regressors and argue that they somewhat improve the predictive efficiency of standard regressors for larger training sets.

Original language	English
Title of host publication	Proceedings of Machine Learning Research
Subtitle of host publication	2026 Conformal and Probabilistic Prediction with Applications
Editors	Johan Hallberg Szabadvary, Ulf Johansson, Henrik Bostrom, Alberto Carlevaro
Volume	329
Publication status	Accepted/In press - 2026

Keywords

auto-calibration
inductive Venn-Abers regressor
cross Venn-Abers regressor

Cite this

@inbook{40359d24a7b3421cae4d601e5c6cf349,

title = "Inductive Venn–Abers and related regressors",

abstract = "Venn–Abers predictors are probabilistic predictors that enjoy appealing properties of validity, but their major limitation is that they are applicable only to the case of binary classification, with a recent extension to bounded regression. We generalize them to the case of unbounded regression, which requires adding an element of conformal prediction. In our simulation and empirical studies we investigate the predictive efficiency of point regressors derived from Venn–Abers regressors and argue that they somewhat improve the predictive efficiency of standard regressors for larger training sets.",

keywords = "auto-calibration, inductive Venn-Abers regressor, cross Venn-Abers regressor",

author = "Vladimir Vovk and Ivan Petej",

year = "2026",

language = "English",

volume = "329",

editor = "\{Hallberg Szabadvary\}, Johan and Ulf Johansson and Henrik Bostrom and Alberto Carlevaro",

booktitle = "Proceedings of Machine Learning Research",

}

Inductive Venn–Abers and related regressors. / Vovk, Vladimir; Petej, Ivan.
Proceedings of Machine Learning Research: 2026 Conformal and Probabilistic Prediction with Applications. ed. / Johan Hallberg Szabadvary; Ulf Johansson; Henrik Bostrom; Alberto Carlevaro. Vol. 329 2026.

Research output: Chapter in Book/Report/Conference proceeding › Chapter (peer-reviewed) › peer-review

TY - CHAP

T1 - Inductive Venn–Abers and related regressors

AU - Vovk, Vladimir

AU - Petej, Ivan

PY - 2026

Y1 - 2026

N2 - Venn–Abers predictors are probabilistic predictors that enjoy appealing properties of validity, but their major limitation is that they are applicable only to the case of binary classification, with a recent extension to bounded regression. We generalize them to the case of unbounded regression, which requires adding an element of conformal prediction. In our simulation and empirical studies we investigate the predictive efficiency of point regressors derived from Venn–Abers regressors and argue that they somewhat improve the predictive efficiency of standard regressors for larger training sets.

AB - Venn–Abers predictors are probabilistic predictors that enjoy appealing properties of validity, but their major limitation is that they are applicable only to the case of binary classification, with a recent extension to bounded regression. We generalize them to the case of unbounded regression, which requires adding an element of conformal prediction. In our simulation and empirical studies we investigate the predictive efficiency of point regressors derived from Venn–Abers regressors and argue that they somewhat improve the predictive efficiency of standard regressors for larger training sets.

KW - auto-calibration

KW - inductive Venn-Abers regressor

KW - cross Venn-Abers regressor

UR - https://arxiv.org/abs/2605.06646

M3 - Chapter (peer-reviewed)

VL - 329

BT - Proceedings of Machine Learning Research

A2 - Hallberg Szabadvary, Johan

A2 - Johansson, Ulf

A2 - Bostrom, Henrik

A2 - Carlevaro, Alberto

ER -

Inductive Venn–Abers and related regressors

Abstract

Keywords

Other files and links

Cite this