强混合样本下单参数指数族参数的平均化经验贝叶斯估计  

AVERAGE EMPIRICAL BAYES ESTIMATION FOR ONE-PARAMETER EXPONENTIAL FAMILIES UNDER STRONG MIXING SAMPLES

作　　者：王甲巍 刘禄勤[1] WANG Jia-wei;LIU Lu-qin(School of Mathematics and Statistics,Wuhan University,Wuhan 430072,China)

机构地区：[1]武汉大学数学与统计学院,湖北武汉430072

出　　处：《数学杂志》2021年第4期365-376,共12页Journal of Mathematics

基　　金：国家自然科学基金资助(11171263).

摘　　要：本文在强混合样本下,利用平均化的思想,研究了一类单参数指数族参数的经验贝叶斯估计,在一定假设条件下得到了该经验贝叶斯估计收敛速度.推广了之前文献的结果,同时,在β混合样本下对该经验贝叶斯估计的风险与贝叶斯估计的风险之间的差值进行了数值模拟.In this paper,the Empirical Bayes(EB)Estimator is investigated by using an average method under strong mixing samples.Meanwhile the convergence rates of proposed EB estimators are obtained under suitable conditions,which generalize some results in literature.The simulation result shows the performance of EB Estimation under β mixing samples.

关 键 词：单参数指数族 经验贝叶斯估计 递归核估计 强混合 平均化 

分 类 号：O212.1[理学—概率论与数理统计]