Pseudo maximum likelihood estimation for the dirichlet-multinomial distribution

Christy Chuang, Christopher Cox

Research output: Contribution to journalArticle

Abstract

Pseudo maximum likelihood estimation (PML) for the Dirichlet-multinomial distribution is proposed and examined in this paper. The procedure is compared to that based on moments (MM) for its asymptotic relative efficiency (ARE) relative to the maximum likelihood estimate (ML). It is found that PML, requiring much less computational effort than ML and possessing considerably higher ARE than MM, constitutes a good compromise between ML and MM. PML is also found to have very high ARE when an estimate for the scale parameter in the Drichlet-multinomial distribution is all that is needed.

Original languageEnglish (US)
Pages (from-to)2293-2311
Number of pages19
JournalCommunications in Statistics - Theory and Methods
Volume14
Issue number10
DOIs
StatePublished - Jan 1 1985
Externally publishedYes

Fingerprint

Pseudo-maximum Likelihood
Dirichlet Distribution
Multinomial Distribution
Asymptotic Relative Efficiency
Maximum Likelihood Estimate
Maximum Likelihood Estimation
Moment
Scale Parameter
Estimate

Keywords

  • asymptotic relative efficiency
  • Dirichlet-multinomial distribution
  • maximum likelihood estimation
  • moment estimation
  • pseudo maximum likelihood estimation

ASJC Scopus subject areas

  • Statistics and Probability

Cite this

Pseudo maximum likelihood estimation for the dirichlet-multinomial distribution. / Chuang, Christy; Cox, Christopher.

In: Communications in Statistics - Theory and Methods, Vol. 14, No. 10, 01.01.1985, p. 2293-2311.

Research output: Contribution to journalArticle

@article{309987c5e31d4481aa3725a25449eab1,
title = "Pseudo maximum likelihood estimation for the dirichlet-multinomial distribution",
abstract = "Pseudo maximum likelihood estimation (PML) for the Dirichlet-multinomial distribution is proposed and examined in this paper. The procedure is compared to that based on moments (MM) for its asymptotic relative efficiency (ARE) relative to the maximum likelihood estimate (ML). It is found that PML, requiring much less computational effort than ML and possessing considerably higher ARE than MM, constitutes a good compromise between ML and MM. PML is also found to have very high ARE when an estimate for the scale parameter in the Drichlet-multinomial distribution is all that is needed.",
keywords = "asymptotic relative efficiency, Dirichlet-multinomial distribution, maximum likelihood estimation, moment estimation, pseudo maximum likelihood estimation",
author = "Christy Chuang and Christopher Cox",
year = "1985",
month = "1",
day = "1",
doi = "10.1080/03610928508829045",
language = "English (US)",
volume = "14",
pages = "2293--2311",
journal = "Communications in Statistics - Theory and Methods",
issn = "0361-0926",
publisher = "Taylor and Francis Ltd.",
number = "10",

}

TY - JOUR

T1 - Pseudo maximum likelihood estimation for the dirichlet-multinomial distribution

AU - Chuang, Christy

AU - Cox, Christopher

PY - 1985/1/1

Y1 - 1985/1/1

N2 - Pseudo maximum likelihood estimation (PML) for the Dirichlet-multinomial distribution is proposed and examined in this paper. The procedure is compared to that based on moments (MM) for its asymptotic relative efficiency (ARE) relative to the maximum likelihood estimate (ML). It is found that PML, requiring much less computational effort than ML and possessing considerably higher ARE than MM, constitutes a good compromise between ML and MM. PML is also found to have very high ARE when an estimate for the scale parameter in the Drichlet-multinomial distribution is all that is needed.

AB - Pseudo maximum likelihood estimation (PML) for the Dirichlet-multinomial distribution is proposed and examined in this paper. The procedure is compared to that based on moments (MM) for its asymptotic relative efficiency (ARE) relative to the maximum likelihood estimate (ML). It is found that PML, requiring much less computational effort than ML and possessing considerably higher ARE than MM, constitutes a good compromise between ML and MM. PML is also found to have very high ARE when an estimate for the scale parameter in the Drichlet-multinomial distribution is all that is needed.

KW - asymptotic relative efficiency

KW - Dirichlet-multinomial distribution

KW - maximum likelihood estimation

KW - moment estimation

KW - pseudo maximum likelihood estimation

UR - http://www.scopus.com/inward/record.url?scp=0006748350&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0006748350&partnerID=8YFLogxK

U2 - 10.1080/03610928508829045

DO - 10.1080/03610928508829045

M3 - Article

AN - SCOPUS:0006748350

VL - 14

SP - 2293

EP - 2311

JO - Communications in Statistics - Theory and Methods

JF - Communications in Statistics - Theory and Methods

SN - 0361-0926

IS - 10

ER -