欢迎访问中国科学院大学学报,今天是

中国科学院大学学报 ›› 2007, Vol. 24 ›› Issue (1): 9-17.DOI: 10.7523/j.issn.2095-6134.2007.1.002

• 论文 • 上一篇    下一篇

刻度指数族参数的经验Bayes检验的收敛速度

魏 莉 孔胜春 韦来生   

  1. 中国科学技术大学统计与金融系
  • 收稿日期:1900-01-01 修回日期:1900-01-01 发布日期:2007-01-15

The Convergence Rates for Empirical Bayes Test of the Parameter in Scale Exponential Family

WEI Li,KONG Sheng-Chun, WEI Lai-Sheng   

  1. Department of Statistics and Finance, USTC

  • Received:1900-01-01 Revised:1900-01-01 Published:2007-01-15

摘要: 在“加权线性损失”下讨论刻度指数族中参数的经验Bayes(EB)检验问题.利用基于 Bessel函数的核估计方法构造了EB检验函数.在适当的条件下证明了获得的EB检验函数是渐近最优的具有收敛速度O(n-1ln6 n).最后给出一个满足定理条件的例子.

关键词: 刻度指数族, 经验Bayes检验, Bessel函数, 收敛速度

Abstract:

In this paper,the empirica lBayes(EB) test problem of the parameter forscale exponential family under weighed linear loss functionsis discussed.The EB tes trules isconstructed by the kernel estimation method based on
Bessel functions.Undersuitable conditionsthe proposed EB test rules are a symptotically optimal with the convergence rate O(n-1(lnn)6).Finally,an example satisfying the conditions of the theorem is shown.

Key words: Scale exponential family, empirical Bayes test, Bessel function, convergence rate

中图分类号: