缺失数据下指数族分布参数的经验Bayes检验

Empirical Bayes Test for Exponential Distrubition Family with Compicated Data

文章在加权线性损失下讨论了带有非随机缺失数据连续型单参数指数族分布参数的经验Bayes检验问题.利用非参数插补的方法,重新构造了参数的经验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.

In this paper,by nonparametric imputation in the case of missing data at non-random,the empirical Bayes test rule for the continuous one-parameter exponential family with missing data is con-structed under weighted linear loss function.The asymptotically optimal property and convergence rates for the proposes EB test are obtained under suitable conditions.