Consistency of the BIC order estimator

I. Csiszár, Paul C. Shields

Research output: Contribution to journalArticle

2 Citations (Scopus)

Abstract

We announce two results on the problem of estimating the order of a Markov chain from observation of a sample path. First is that the Bayesian Information Criterion (BIC) leads to an almost surely consistent estimator. Second is that the Bayesian minimum description length estimator, of which the BIC estimator is an approximation, fails to be consistent for the uniformly distributed i.i.d. process. A key tool is a strong ratio-typicality result for empirical k-block distributions. Complete proofs are given in the authors' article to appear in The Annals of Statistics.

Original languageEnglish
Pages (from-to)123-127
Number of pages5
JournalElectronic Research Announcements of the American Mathematical Society
Volume5
Issue number17
DOIs
Publication statusPublished - Oct 19 1999

Fingerprint

Bayesian Information Criterion
Estimator
Consistent Estimator
Sample Path
Markov chain
Statistics
Approximation
Observation

Keywords

  • Bayesian information criterion
  • Markov chains
  • Order estimation
  • Ratio-typicality

ASJC Scopus subject areas

  • Mathematics(all)

Cite this

Consistency of the BIC order estimator. / Csiszár, I.; Shields, Paul C.

In: Electronic Research Announcements of the American Mathematical Society, Vol. 5, No. 17, 19.10.1999, p. 123-127.

Research output: Contribution to journalArticle

@article{184e31713f9c40c996849042379d21cd,
title = "Consistency of the BIC order estimator",
abstract = "We announce two results on the problem of estimating the order of a Markov chain from observation of a sample path. First is that the Bayesian Information Criterion (BIC) leads to an almost surely consistent estimator. Second is that the Bayesian minimum description length estimator, of which the BIC estimator is an approximation, fails to be consistent for the uniformly distributed i.i.d. process. A key tool is a strong ratio-typicality result for empirical k-block distributions. Complete proofs are given in the authors' article to appear in The Annals of Statistics.",
keywords = "Bayesian information criterion, Markov chains, Order estimation, Ratio-typicality",
author = "I. Csisz{\'a}r and Shields, {Paul C.}",
year = "1999",
month = "10",
day = "19",
doi = "10.1090/S1079-6762-99-00070-0",
language = "English",
volume = "5",
pages = "123--127",
journal = "Electronic Research Announcements in Mathematical Sciences",
issn = "1935-9179",
publisher = "American Mathematical Society",
number = "17",

}

TY - JOUR

T1 - Consistency of the BIC order estimator

AU - Csiszár, I.

AU - Shields, Paul C.

PY - 1999/10/19

Y1 - 1999/10/19

N2 - We announce two results on the problem of estimating the order of a Markov chain from observation of a sample path. First is that the Bayesian Information Criterion (BIC) leads to an almost surely consistent estimator. Second is that the Bayesian minimum description length estimator, of which the BIC estimator is an approximation, fails to be consistent for the uniformly distributed i.i.d. process. A key tool is a strong ratio-typicality result for empirical k-block distributions. Complete proofs are given in the authors' article to appear in The Annals of Statistics.

AB - We announce two results on the problem of estimating the order of a Markov chain from observation of a sample path. First is that the Bayesian Information Criterion (BIC) leads to an almost surely consistent estimator. Second is that the Bayesian minimum description length estimator, of which the BIC estimator is an approximation, fails to be consistent for the uniformly distributed i.i.d. process. A key tool is a strong ratio-typicality result for empirical k-block distributions. Complete proofs are given in the authors' article to appear in The Annals of Statistics.

KW - Bayesian information criterion

KW - Markov chains

KW - Order estimation

KW - Ratio-typicality

UR - http://www.scopus.com/inward/record.url?scp=85009813165&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85009813165&partnerID=8YFLogxK

U2 - 10.1090/S1079-6762-99-00070-0

DO - 10.1090/S1079-6762-99-00070-0

M3 - Article

VL - 5

SP - 123

EP - 127

JO - Electronic Research Announcements in Mathematical Sciences

JF - Electronic Research Announcements in Mathematical Sciences

SN - 1935-9179

IS - 17

ER -