华东师范大学学报(自然科学版) ›› 2004, Vol. 2004 ›› Issue (3): 42-47,1.

• 数学 统计学 • 上一篇    下一篇

关于两种贝叶斯模型的一致性比较

濮晓龙， 徐靖   

  1. 华东师范大学 统计系，上海 200062
  • 收稿日期:2003-12-16 修回日期:2004-01-16 出版日期:2004-09-25 发布日期:2004-09-25
  • 通讯作者: 濮晓龙

Comparison of Two Bayesian Models

PU Xiao-long, XU Jing

  

  1. Department of Statistics, East China Normal University, Shanghai 200062，China
  • Received:2003-12-16 Revised:2004-01-16 Online:2004-09-25 Published:2004-09-25
  • Contact: PU Xiao-long

摘要： 主要讨论两样本情况下两种不同贝叶斯估计方法之间存在的差异.方法一是假定xi（i=1,2）关于θi（i=1,2）条件独立；方法二假定x1、x2无条件独立.在考虑θ1的统计分析时，研究这二种方法的结果有无明显的差别.由实际分析可以得出结论：用两种模型对θ1进行贝叶斯估计差异不大.

关键词: 贝叶斯模型, 条件独立, 无条件独立, 贝叶斯估计, 贝叶斯模型, 条件独立, 无条件独立, 贝叶斯估计

Abstract: This article discusses the difference between two Bayesian models. With model 1, it is assumed that xi（i=1,2）is conditional independent under θi（i=1,2）. While with model 2, it is assumed that xi（i=1,2）is unconditional independent. The question is that when working for the Bayesian estimation of θ1, is there any difference on the results under the two different models? The numerical comparison shows that there is only small difference between the results. So the simpler model 2 can replace the model 1, when Bayesian estimation of θ1 is calculated.

Key words: conditional independent, unconditional independent, Bayesian estimation, Bayesian model, conditional independent, unconditional independent, Bayesian estimation

中图分类号: