The aim in this article is to provide a means to undertake Bayesian inference for mixture models when the likelihood function is raised to a power between 0 and 1. The main purpose for doing this is to guarantee a strongly consistent model and hence, make it possible to compare the consistent posterior with the correct posterior, looking for signs of discrepancy. This will be explained in detail in the article. Another purpose would be for simulated annealing algorithms. In particular, for the widely used mixture of Dirichlet process model, it is far from obvious how to undertake inference via Markov chain Monte Carlo methods when the likelihood is raised to a power other than 1. In this article, we demonstrate how posterior sampling can be carried out when using a power likelihood. Matlab code to implement the algorithm is available as supplementary material. © 2013 American Statistical Association, Institute of Mathematical Statistics, and Interface Foundation of North America.

Bayesian nonparametric inference for the power likelihood

ANTONIANO VILLALOBOS, Isadora;
2013

Abstract

The aim in this article is to provide a means to undertake Bayesian inference for mixture models when the likelihood function is raised to a power between 0 and 1. The main purpose for doing this is to guarantee a strongly consistent model and hence, make it possible to compare the consistent posterior with the correct posterior, looking for signs of discrepancy. This will be explained in detail in the article. Another purpose would be for simulated annealing algorithms. In particular, for the widely used mixture of Dirichlet process model, it is far from obvious how to undertake inference via Markov chain Monte Carlo methods when the likelihood is raised to a power other than 1. In this article, we demonstrate how posterior sampling can be carried out when using a power likelihood. Matlab code to implement the algorithm is available as supplementary material. © 2013 American Statistical Association, Institute of Mathematical Statistics, and Interface Foundation of North America.
File in questo prodotto:
File Dimensione Formato  
mdpc_v4.pdf

non disponibili

Dimensione 314.84 kB
Formato Adobe PDF
314.84 kB Adobe PDF   Visualizza/Apri
Antoniano-Walker(2012)Power likelihood(Final).pdf

non disponibili

Dimensione 306.26 kB
Formato Adobe PDF
306.26 kB Adobe PDF   Visualizza/Apri

I documenti in ARCA sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: http://hdl.handle.net/10278/3711720
Citazioni
  • ???jsp.display-item.citation.pmc??? ND
  • Scopus 4
  • ???jsp.display-item.citation.isi??? 3
social impact