首页> 外文期刊>Statistical Methods and Applications >Discussion of 'Sampling schemes for generalized DP random effects models'
【24h】

Discussion of 'Sampling schemes for generalized DP random effects models'

机译:关于“广义DP随机效应模型的采样方案”的讨论

获取原文
获取原文并翻译 | 示例
       

摘要

I congratulate the authors on a stimulating and interesting paper. They take a refreshingly different perspective of DP mixture models and see different aspects from what is traditionally discussed in the non-parametric Bayesian literature. I would like to highlight some features of the discussed models and implementations that are stated in the paper, but might not be noticed by a casual reader who is not familiar with these models. First, the paper introduces the GLMDM as a special case of the popular DP mixture models, but then proceeds discussing it (correctly, of course) as a semi-parametric extension of GLMMs. The posterior MCMC for the random partition A, the precision m and other features related to the DP model do not exploit the nature of the mixing kernel as a GLM.
机译:我祝贺作者发表了一篇有趣而有趣的论文。他们对DP混合模型采用了令人耳目一新的不同观点,并且从非参数贝叶斯文献中传统讨论的内容中看到了不同的方面。我想强调本文中讨论的模型和实现的某些功能,但那些不熟悉这些模型的随便的读者可能不会注意到。首先,本文介绍了GLMDM作为流行的DP混合模型的特例,但随后(当然正确)将其作为GLMM的半参数扩展进行了讨论。随机分区A的后MCMC,精度m和与DP模型相关的其他特征未利用混合内核作为GLM的性质。

著录项

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱:kefu@zhangqiaokeyan.com

京公网安备:11010802029741号 ICP备案号:京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有
  • 客服微信

  • 服务号