Scholarlink

We propose a data augmented Bayesian estimation method for rectangular hidden Markov models, which represent a variation of hidden Markov models where the number of hidden states is allowed to vary across time. The method involves augmenting the target posterior distribution with the hidden variables, resulting in closed-form expressions for updating the posterior parameters during the Gibbs sampling procedure. Furthermore, the implementation of the reversible jump Markov chain Monte Carlo algorithm enables selection of the number of hidden states at each time occasion.

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

A Data Augmented Bayesian Approach for Rectangular Hidden Markov Models

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Romina Peruilh Bagolini,
  • Daniele Tancini,
  • Francesco Bartolucci,
  • Silvia Pandolfi

摘要

We propose a data augmented Bayesian estimation method for rectangular hidden Markov models, which represent a variation of hidden Markov models where the number of hidden states is allowed to vary across time. The method involves augmenting the target posterior distribution with the hidden variables, resulting in closed-form expressions for updating the posterior parameters during the Gibbs sampling procedure. Furthermore, the implementation of the reversible jump Markov chain Monte Carlo algorithm enables selection of the number of hidden states at each time occasion.