伽玛分布族参数的经验Bayes双边检验的收敛速度:NA样本情形

CONVERGENCE RATES OF EMPIRICAL BAYES TWO-SIDED TEST FOR THE PARAMETER OF GAMMA DISTRIBUTION FAMILIES∶IN THE CASE OF NA SAMPLES

本文研究了NA样本情形下,伽玛分布族形状参数的经验Bayes(EB)双边检验问题.利用概率密度函数的核估计,构造了参数的经验Bayes检验函数,并在适当的条件下,证明了所提出的经验Bayes检验函数的渐近取优(a.o.)性,获得了其收敛速度.

In this paper.the empirical Bayes（EB） two-sided test problem of shape parameter for Gamma distribution families in the case of identically distributed and negatively associated （NA） samples is investigated. By using the kernel-type density estimation, the empirical Bayes two-sided test rules are constructed.the asymptotically optimal property and convergence rates for the proposed EB testrules are obtained under suitable conditions.