右删失数据下Pareto分布参数的经验Bayes渐近性质  

Asymptotical Property of an Empirical Bayes Test for the Parameter of Pareto Distribution under Right Censored Data

作　　者：黄娟[1] 

机构地区：[1]广东海洋大学理学院,湛江524088

出　　处：《工程数学学报》2010年第5期781-788,共8页Chinese Journal of Engineering Mathematics

基　　金：广东海洋大学科研资助项目(1012138;0612163)~~

摘　　要：本文借助经验Bayes方法研究了Pareto分布参数的双边检验,由于此方法无需对先验分布强加限制,因而非常稳健。在随机右删失的情形下,为了估计未知的边缘密度函数及其导数,首先引入核估计方法,从而,进一步构造了参数的经验Bayes检验函数,并获得了所提出的检验函数的渐近最优性及其收敛速度。In this paper,we study the two-sided test of Pareto distribution by the empirical Bayes approach.Nonparametric empirical Bayes is more robust since it imposes no restriction on the prior.In order to estimate an unknown marginal density and its derivative,a kernel method is firstly introduced. Then the test rule for the parameter of Pareto distribution is constructed in case of randomly censored from the right.The asymptotically optimal property and convergence rates for the proposed EB test are derived.

关 键 词：右删失数据 经验BAYES 渐近最优性 收敛速度 

分 类 号：O212.1[理学—概率论与数理统计]