Geometry of goodness-of-fit testing in high dimensional low sample size modelling

Paul Marriott, Radka Sabolova, Germain van Bever, Frank Critchley

Research output: Contribution in Book/Catalog/Report/Conference proceedingConference contribution

Abstract

We introduce a new approach to goodness-of-fit testing in the high dimensional, sparse extended multinomial context. The paper takes a computational information geometric approach, extending classical higher order asymptotic theory. We show why the Wald – equivalently, the Pearson χ2 and score statistics – are unworkable in this context, but that the deviance has a simple, accurate and tractable sampling distribution even for moderate sample sizes. Issues of uniformity of asymptotic approximations across model space are discussed. A variety of important applications and extensions are noted.

Original languageEnglish
Title of host publicationGeometric Science of Information - 2nd International Conference, GSI 2015, Proceedings
EditorsFrank Nielsen, Frank Nielsen, Frank Nielsen, Frederic Barbaresco, Frederic Barbaresco, Frank Nielsen
PublisherSpringer Verlag
Pages569-576
Number of pages8
ISBN (Print)9783319250397, 9783319250397
DOIs
Publication statusPublished - 2015
Externally publishedYes
Event2nd International Conference on Geometric Science of Information, GSI 2015 - Palaiseau, France
Duration: 28 Oct 201530 Oct 2015

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume9389
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference2nd International Conference on Geometric Science of Information, GSI 2015
Country/TerritoryFrance
CityPalaiseau
Period28/10/1530/10/15

Fingerprint

Dive into the research topics of 'Geometry of goodness-of-fit testing in high dimensional low sample size modelling'. Together they form a unique fingerprint.

Cite this