Bayesian Analysis

Splitting and merging components of a nonconjugate Dirichlet process mixture model

Sonia Jain and Radford M. Neal

Full-text: Open access

Abstract

The inferential problem of associating data to mixture components is difficult when components are nearby or overlapping. We introduce a new split-merge Markov chain Monte Carlo technique that efficiently classifies observations by splitting and merging mixture components of a nonconjugate Dirichlet process mixture model. Our method, which is a Metropolis-Hastings procedure with split-merge proposals, samples clusters of observations simultaneously rather than incrementally assigning observations to mixture components. Split-merge moves are produced by exploiting properties of a restricted Gibbs sampling scan. A simulation study compares the new split-merge technique to a nonconjugate version of Gibbs sampling and an incremental Metropolis-Hastings technique. The results demonstrate the improved performance of the new sampler.

Article information

Source
Bayesian Anal., Volume 2, Number 3 (2007), 445-472.

Dates
First available in Project Euclid: 22 June 2012

Permanent link to this document
https://projecteuclid.org/euclid.ba/1340370720

Digital Object Identifier
doi:10.1214/07-BA219

Mathematical Reviews number (MathSciNet)
MR2342168

Zentralblatt MATH identifier
1331.62145

Keywords
Bayesian model Markov chain Monte Carlo split-merge moves nonconjugate prior

Citation

Jain, Sonia; Neal, Radford M. Splitting and merging components of a nonconjugate Dirichlet process mixture model. Bayesian Anal. 2 (2007), no. 3, 445--472. doi:10.1214/07-BA219. https://projecteuclid.org/euclid.ba/1340370720


Export citation

See also

  • Related item: David B. Dahl. Comment on article by Jain and Neal. Bayesian Anal., Vol. 2, Iss. 3 (2007), 473-477.
  • Related item: C. P. Robert. Comment on article by Jain and Neal. Bayesian Anal., Vol. 2, Iss. 3 (2007), 479-482.
  • Related item: Steven N. MacEachern. Comment on article by Jain and Neal. Bayesian Anal., Vol. 2, Iss. 3 (2007), 483-494.
  • Related item: Sonia Jain, Radford M. Neal. Rejoinder. Bayesian Anal., Vol. 2, Iss. 3 (2007), 495-500.