Gibbs ensembles for nearly compatible and incompatible conditional models

Shyh Huei Chen, Edward H. Ip, Yuchung J. Wang

Research output: Contribution to journalArticle

11 Scopus citations

Abstract

The Gibbs sampler has been used exclusively for compatible conditionals that converge to a unique invariant joint distribution. However, conditional models are not always compatible. In this paper, a Gibbs sampling-based approachusing the Gibbs ensembleis proposed for searching for a joint distribution that deviates least from a prescribed set of conditional distributions. The algorithm can be easily scalable, such that it can handle large data sets of high dimensionality. Using simulated data, we show that the proposed approach provides joint distributions that are less discrepant from the incompatible conditionals than those obtained by other methods discussed in the literature. The ensemble approach is also applied to a data set relating to geno-polymorphism and response to chemotherapy for patients with metastatic colorectal cancer.

Original languageEnglish (US)
Pages (from-to)1760-1769
Number of pages10
JournalComputational Statistics and Data Analysis
Volume55
Issue number4
DOIs
StatePublished - Apr 1 2011

All Science Journal Classification (ASJC) codes

  • Statistics and Probability
  • Computational Mathematics
  • Computational Theory and Mathematics
  • Applied Mathematics

Keywords

  • Conditionally specified distribution
  • Ensemble method
  • Gibbs sampler
  • Linear programming
  • Odds ratio

Fingerprint Dive into the research topics of 'Gibbs ensembles for nearly compatible and incompatible conditional models'. Together they form a unique fingerprint.

  • Cite this