中文摘要：

在加权损失函数下讨论了刻度指数族中参数的经验Bayes（EB）双边检验问题．利用概率密度函数及其导数的核估计方法构造了EB检验函数并证明了其渐近最优性，获得了其收敛速度．最后，给出了一个符合定理条件的例子．

英文摘要：

Under the weighed loss function, the two-sided empirical Bayes （EB） test of parameters for the scale exponential family was discussed. The EB test rules were constructed by kernel estimation method. The asymptotical optimality and convergence rates of the EB test rules were obtained. Finally, an examplesatisfying the conditions of the theorem was shown.