Universal predictive systems. / Vovk, Vladimir.

In: Pattern Recognition, 05.05.2021.

Research output: Contribution to journalArticlepeer-review

Forthcoming

Standard

Universal predictive systems. / Vovk, Vladimir.

In: Pattern Recognition, 05.05.2021.

Research output: Contribution to journalArticlepeer-review

Harvard

Vovk, V 2021, 'Universal predictive systems', Pattern Recognition.

APA

Vovk, V. (Accepted/In press). Universal predictive systems. Pattern Recognition.

Vancouver

Vovk V. Universal predictive systems. Pattern Recognition. 2021 May 5.

Author

Vovk, Vladimir. / Universal predictive systems. In: Pattern Recognition. 2021.

BibTeX

@article{fbd4493806134488afbdb7369e9cf5ec,
title = "Universal predictive systems",
abstract = "This paper describes probability forecasting systems that are universal, or universally consistent, in the sense of being consistent under any data-generating distribution, assuming that the observations are produced independently in the IID fashion. The notion of universal consistency is asymptotic and does not imply any small-sample guarantees of validity. On the other hand, the method of conformal prediction has been recently adapted to producing predictive distributions that satisfy a natural property of small-sample validity, namely they are automatically probabilistically calibrated. The main result of the paper is the existence of universal conformal predictive systems, which output predictive distributions that are both probabilistically calibrated and universally consistent.",
keywords = "conformal prediction, predictive distribution, probabilistic calibration, universal consistency",
author = "Vladimir Vovk",
year = "2021",
month = may,
day = "5",
language = "English",
journal = "Pattern Recognition",
issn = "0031-3203",
publisher = "Elsevier Limited",

}

RIS

TY - JOUR

T1 - Universal predictive systems

AU - Vovk, Vladimir

PY - 2021/5/5

Y1 - 2021/5/5

N2 - This paper describes probability forecasting systems that are universal, or universally consistent, in the sense of being consistent under any data-generating distribution, assuming that the observations are produced independently in the IID fashion. The notion of universal consistency is asymptotic and does not imply any small-sample guarantees of validity. On the other hand, the method of conformal prediction has been recently adapted to producing predictive distributions that satisfy a natural property of small-sample validity, namely they are automatically probabilistically calibrated. The main result of the paper is the existence of universal conformal predictive systems, which output predictive distributions that are both probabilistically calibrated and universally consistent.

AB - This paper describes probability forecasting systems that are universal, or universally consistent, in the sense of being consistent under any data-generating distribution, assuming that the observations are produced independently in the IID fashion. The notion of universal consistency is asymptotic and does not imply any small-sample guarantees of validity. On the other hand, the method of conformal prediction has been recently adapted to producing predictive distributions that satisfy a natural property of small-sample validity, namely they are automatically probabilistically calibrated. The main result of the paper is the existence of universal conformal predictive systems, which output predictive distributions that are both probabilistically calibrated and universally consistent.

KW - conformal prediction

KW - predictive distribution

KW - probabilistic calibration

KW - universal consistency

M3 - Article

JO - Pattern Recognition

JF - Pattern Recognition

SN - 0031-3203

ER -