E-values: Calibration, combination, and applications

Forthcoming in the Annals of Statistics

48 Pages Posted: 1 Jan 2020 Last revised: 22 Sep 2020

See all articles by Vladimir Vovk

Vladimir Vovk

Ruodu Wang

University of Waterloo - Department of Statistics and Actuarial Science

Date Written: December 14, 2019

Abstract

Multiple testing of a single hypothesis and testing multiple hypotheses are usually done in terms of p-values. In this paper we replace p-values with their natural competitor, e-values, which are closely related to betting, Bayes factors, and likelihood ratios. We demonstrate that e-values are often mathematically more tractable; in particular, in multiple testing of a single hypothesis, e-values can be merged simply by averaging them. This allows us to develop ecient procedures using e-values for testing multiple hypotheses.

Keywords: Hypothesis testing, multiple hypothesis testing, Bayes factor, test martingale, admissible decisions

JEL Classification: C12

Suggested Citation

Vovk, Vladimir and Wang, Ruodu, E-values: Calibration, combination, and applications (December 14, 2019). Forthcoming in the Annals of Statistics, Available at SSRN: https://ssrn.com/abstract=3504009 or http://dx.doi.org/10.2139/ssrn.3504009

Ruodu Wang (Contact Author)

University of Waterloo - Department of Statistics and Actuarial Science ( email )

Waterloo, Ontario N2L 3G1
Canada

No contact information is available for Vladimir Vovk

Here is the Coronavirus
related research on SSRN

Paper statistics

Downloads
102
Abstract Views
421
rank
288,338
PlumX Metrics