Testing Equivalence of Multinomial Distributions

Statistics & Probability Letters, Volume 124, May 2017, Pages 77–82.

10 Pages Posted: 10 Sep 2021

Date Written: May 12, 2016

Abstract

We introduce the smooth total variation distance as a natural choice to establish equivalence. We propose asymptotic and bootstrap-based tests for which asymptotic optimality is shown. The finite sample performance is studied by simulations. Then we apply tests to real data sets. The tests are implemented in VB.net and available on GitHub, see SourceCode (2018).

Keywords: Equivalence, Testing, Multinomial, Bootstrap, Optimality, Benford's Law, Mendelian Inheritance, Decimal Digits of Pi

JEL Classification: C12

Suggested Citation

Ostrovski, Vladimir, Testing Equivalence of Multinomial Distributions (May 12, 2016). Statistics & Probability Letters, Volume 124, May 2017, Pages 77–82., Available at SSRN: https://ssrn.com/abstract=3241589 or http://dx.doi.org/10.2139/ssrn.3241589

Vladimir Ostrovski (Contact Author)

Talanx Asset Management ( email )

Charles-de-Gaulle-Platz 1
Cologne, DE NRW 50679
Germany

Do you have a job opening that you would like to promote on SSRN?

Paper statistics

Downloads
21
Abstract Views
190
PlumX Metrics