Comparing Bayesian and Likelihood Approaches to Parameter Estimation of a Negative Binomial Distribution

Kód výsledku v IS VaVaI

<a href="https://www.isvavai.cz/riv?ss=detail&h=RIV%2F60162694%3AG42__%2F10%3A00427046" target="_blank" >RIV/60162694:G42__/10:00427046 - isvavai.cz</a>

Nalezeny alternativní kódy

RIV/00216305:26210/10:PU87475

Výsledek na webu

—

DOI - Digital Object Identifier

—

Jazyk výsledku

angličtina

Název v původním jazyce

Comparing Bayesian and Likelihood Approaches to Parameter Estimation of a Negative Binomial Distribution

Popis výsledku v původním jazyce

It is well known that the maximum-likelihood estimates of a negative binomial distribution are biased and their numerical calculation tends to be error-prone for some extremal parameter values. In this case, it is necessary to find a different approach.A good approach seems to be a Bayesian one. The paper aims to compare both approaches. Using simulations it is shown that a Bayesian approach produces good results even in cases when the likelihood one fails. Moreover, estimators obtained using the Bayesian approach seems to be as good as bias corrected likelihood estimators as well as the numerically simpler quasi-likelihood estimators. Finally, based on the simulation results a recommendation is formulated that leads to obtaining optimal estimators for arbitrary parameter values. A system of Matlab programmes for calculating the estimates in question can be obtained from the first author.

Název v anglickém jazyce

Comparing Bayesian and Likelihood Approaches to Parameter Estimation of a Negative Binomial Distribution

Popis výsledku anglicky

It is well known that the maximum-likelihood estimates of a negative binomial distribution are biased and their numerical calculation tends to be error-prone for some extremal parameter values. In this case, it is necessary to find a different approach.A good approach seems to be a Bayesian one. The paper aims to compare both approaches. Using simulations it is shown that a Bayesian approach produces good results even in cases when the likelihood one fails. Moreover, estimators obtained using the Bayesian approach seems to be as good as bias corrected likelihood estimators as well as the numerically simpler quasi-likelihood estimators. Finally, based on the simulation results a recommendation is formulated that leads to obtaining optimal estimators for arbitrary parameter values. A system of Matlab programmes for calculating the estimates in question can be obtained from the first author.

Druh

D - Stať ve sborníku

CEP obor

BB - Aplikovaná statistika, operační výzkum

OECD FORD obor

—

Projekt

<a href="/cs/project/1M06047" target="_blank" >1M06047: Centrum pro jakost a spolehlivost výroby</a><br>

Návaznosti

Z - Vyzkumny zamer (s odkazem do CEZ)

Rok uplatnění

2010

Kód důvěrnosti údajů

S - Úplné a pravdivé údaje o projektu nepodléhají ochraně podle zvláštních právních předpisů

Název statě ve sborníku

MENDEL 2010

ISBN

978-80-214-4120-0

ISSN

—

e-ISSN

—

Počet stran výsledku

8

Strana od-do

—

Název nakladatele

University of Technology

Místo vydání

Brno

Místo konání akce

Brno

Datum konání akce

1. 1. 2010

Typ akce podle státní příslušnosti

WRD - Celosvětová akce

Kód UT WoS článku

—