利用经验贝叶斯方法研究了伽玛分布参数的双边检验问题,构造了一个在历史样本被随机右删失的条件下参数的经验Bayes检验函数,在适当的条件下证明了所提出的经验Bayes检验函数的渐近最优性,并获得了它的收敛速度可任意接近O(n~(-1/2)).
This paper studies two-sided test problem of Gamma distribution by the empirical Bayes approach.The test rule for the parameter of Gamma distribution is constructed under the condition that the past samples are randomly censored from the right.The asymptotically optimal property is obtained and it is show that the convergence rate for the proposed EB test is arbitrarily close to 0(n~(-1/2)).