Scholarlink

<p>We propose to use limiting sampling distributions of the method of moments estimator to construct Metropolis-Hastings transition densities in Markov chain Monte Carlo (MCMC) algorithms. The proposed technique is applied to developing an efficient MCMC algorithm for Bayesian estimation of a finite beta mixture model, which demonstrates excellent performance in a Monte-Carlo study. More generally, the technique can be useful for models in which the method of moments (but not the maximum likelihood) estimator for a subset of parameters is available in a closed form; examples include mixtures of gamma and Dirichlet densities.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Mcmc proposals based on method of moments with an application to finite beta mixtures

  • 期刊论文

  • 发表日期：

    2026年1月26日

查看全文
  • Andriy Norets,
  • Xun Tang

摘要

We propose to use limiting sampling distributions of the method of moments estimator to construct Metropolis-Hastings transition densities in Markov chain Monte Carlo (MCMC) algorithms. The proposed technique is applied to developing an efficient MCMC algorithm for Bayesian estimation of a finite beta mixture model, which demonstrates excellent performance in a Monte-Carlo study. More generally, the technique can be useful for models in which the method of moments (but not the maximum likelihood) estimator for a subset of parameters is available in a closed form; examples include mixtures of gamma and Dirichlet densities.