讨论两两NQD序列下线性指数分布参数的经验Bayes(EB)双边检验问题.利用概率密度函数的核估计方法,构造参数的EB检验函数,在适当的条件下证明EB检验函数是渐近最优的,并获得它的收敛速度.举出一个满足定理条件的例子.
The problem of empirical Bayes (EB) two-sided test of linear exponential distributive parame- ter with NQD random series in twos was discussed. By using kernel estimation method of probability den- sity function, the EB test function of the parameter was constructed. It was proofed under some suitable conditions that the EB test function was asymptotically optimal and its convergence rate was obtained. Fi- nally, an example satisfying the conditions of theorem was given.