Information criterion for Gaussian change-point model

Yoshiyuki Ninomiya

Research output: Contribution to journalArticle

24 Citations (Scopus)

Abstract

AIC-type information criterion is generally estimated by the bias-corrected maximum log-likelihood. In regular models, the bias can be estimated by p, where p is the number of parameters. The present paper considers the AIC-type information criterion for change-point models which are not regular, the bias of which will not be the same as for regular models. The bias is shown to depend on the expected maximum of a random walk with negative drift. Furthermore, it is shown that by using an approximation to a Brownian motion, the evaluated bias is given by 3m + pm (not m + pm), where m is the number of change-points and pm is the number of regular parameters, which differs from regular models.

Original languageEnglish
Pages (from-to)237-247
Number of pages11
JournalStatistics and Probability Letters
Volume72
Issue number3
DOIs
Publication statusPublished - May 1 2005

Fingerprint

Change-point Model
Information Criterion
Gaussian Model
p.m.
Change Point
Brownian motion
Random walk
Likelihood
Information criterion
Change point
Model
Approximation

All Science Journal Classification (ASJC) codes

  • Statistics and Probability
  • Statistics, Probability and Uncertainty

Cite this

Information criterion for Gaussian change-point model. / Ninomiya, Yoshiyuki.

In: Statistics and Probability Letters, Vol. 72, No. 3, 01.05.2005, p. 237-247.

Research output: Contribution to journalArticle

Ninomiya, Yoshiyuki. / Information criterion for Gaussian change-point model. In: Statistics and Probability Letters. 2005 ; Vol. 72, No. 3. pp. 237-247.
@article{b9abca0eaab9478c962914b84a073b8f,
title = "Information criterion for Gaussian change-point model",
abstract = "AIC-type information criterion is generally estimated by the bias-corrected maximum log-likelihood. In regular models, the bias can be estimated by p, where p is the number of parameters. The present paper considers the AIC-type information criterion for change-point models which are not regular, the bias of which will not be the same as for regular models. The bias is shown to depend on the expected maximum of a random walk with negative drift. Furthermore, it is shown that by using an approximation to a Brownian motion, the evaluated bias is given by 3m + pm (not m + pm), where m is the number of change-points and pm is the number of regular parameters, which differs from regular models.",
author = "Yoshiyuki Ninomiya",
year = "2005",
month = "5",
day = "1",
doi = "10.1016/j.spl.2004.10.037",
language = "English",
volume = "72",
pages = "237--247",
journal = "Statistics and Probability Letters",
issn = "0167-7152",
publisher = "Elsevier",
number = "3",

}

TY - JOUR

T1 - Information criterion for Gaussian change-point model

AU - Ninomiya, Yoshiyuki

PY - 2005/5/1

Y1 - 2005/5/1

N2 - AIC-type information criterion is generally estimated by the bias-corrected maximum log-likelihood. In regular models, the bias can be estimated by p, where p is the number of parameters. The present paper considers the AIC-type information criterion for change-point models which are not regular, the bias of which will not be the same as for regular models. The bias is shown to depend on the expected maximum of a random walk with negative drift. Furthermore, it is shown that by using an approximation to a Brownian motion, the evaluated bias is given by 3m + pm (not m + pm), where m is the number of change-points and pm is the number of regular parameters, which differs from regular models.

AB - AIC-type information criterion is generally estimated by the bias-corrected maximum log-likelihood. In regular models, the bias can be estimated by p, where p is the number of parameters. The present paper considers the AIC-type information criterion for change-point models which are not regular, the bias of which will not be the same as for regular models. The bias is shown to depend on the expected maximum of a random walk with negative drift. Furthermore, it is shown that by using an approximation to a Brownian motion, the evaluated bias is given by 3m + pm (not m + pm), where m is the number of change-points and pm is the number of regular parameters, which differs from regular models.

UR - http://www.scopus.com/inward/record.url?scp=15844402690&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=15844402690&partnerID=8YFLogxK

U2 - 10.1016/j.spl.2004.10.037

DO - 10.1016/j.spl.2004.10.037

M3 - Article

AN - SCOPUS:15844402690

VL - 72

SP - 237

EP - 247

JO - Statistics and Probability Letters

JF - Statistics and Probability Letters

SN - 0167-7152

IS - 3

ER -