Universally consistent conformal predictive distributions

Vladimir Vovk

Universally consistent conformal predictive distributions

Vladimir Vovk

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

Abstract

This paper describes conformal predictive systems that are universally consistent in the sense of being consistent under any data-generating distribution, assuming that the observations are produced independently in the IID fashion. Being conformal, these predictive systems satisfy a natural property of small-sample validity, namely they are automatically calibrated in probability.

Original language	English
Title of host publication	Proceedings of Machine Learning Research
Editors	Alex Gammerman, Vladimir Vovk, Zhiyuan Luo, Evgueni Smirnov
Pages	105-122
Number of pages	18
Volume	105
Publication status	Published - 29 Aug 2019

Keywords

calibration in probability
conformal prediction
predictive distribution
universal consistency

Access to Document

http://proceedings.mlr.press/v105/vovk19a/vovk19a.pdf

Cite this

@inbook{627cf4416ea34f5aaa21bb9bd9ed7780,

title = "Universally consistent conformal predictive distributions",

abstract = "This paper describes conformal predictive systems that are universally consistent in the sense of being consistent under any data-generating distribution, assuming that the observations are produced independently in the IID fashion. Being conformal, these predictive systems satisfy a natural property of small-sample validity, namely they are automatically calibrated in probability.",

keywords = "calibration in probability, conformal prediction, predictive distribution, universal consistency",

author = "Vladimir Vovk",

year = "2019",

month = aug,

day = "29",

language = "English",

volume = "105",

pages = "105--122",

editor = "Alex Gammerman and Vladimir Vovk and Zhiyuan Luo and Evgueni Smirnov",

booktitle = "Proceedings of Machine Learning Research",

}

TY - CHAP

T1 - Universally consistent conformal predictive distributions

AU - Vovk, Vladimir

PY - 2019/8/29

Y1 - 2019/8/29

N2 - This paper describes conformal predictive systems that are universally consistent in the sense of being consistent under any data-generating distribution, assuming that the observations are produced independently in the IID fashion. Being conformal, these predictive systems satisfy a natural property of small-sample validity, namely they are automatically calibrated in probability.

AB - This paper describes conformal predictive systems that are universally consistent in the sense of being consistent under any data-generating distribution, assuming that the observations are produced independently in the IID fashion. Being conformal, these predictive systems satisfy a natural property of small-sample validity, namely they are automatically calibrated in probability.

KW - calibration in probability

KW - conformal prediction

KW - predictive distribution

KW - universal consistency

M3 - Chapter (peer-reviewed)

VL - 105

SP - 105

EP - 122

BT - Proceedings of Machine Learning Research

A2 - Gammerman, Alex

A2 - Vovk, Vladimir

A2 - Luo, Zhiyuan

A2 - Smirnov, Evgueni

ER -