目的讨论负相依样本下一类特殊指数分布族的经验贝叶斯单边检验问题。方法利用概率密度的变窗核估计方法。结果得到了经验贝叶斯检验函数,在适当的条件下获得其收敛速度,并给出了一个满足定理条件的例子。结论该检验函数是渐近最优的。
Aim To study the Bayes test problem for a special exponential family under negatively associated sampies. Methods Density estimation by using the kernel estimation with variable window width. Results The asymptotic optimality and convergence rates of the empirical Bayes test function are obtained under some suitable conditions, an example about main results is given. Conclusion The empirical Bayes test function is constructed.