在单参数双边截断型分布族中参数估计的一种新的渐近效率  

A New Kind of Asympototic Efficiency in One-Parameter Two-Sided Truncated Distribution Family

作　　者：陈桂景[1] 

机构地区：[1]安徽大学数学系,安徽合肥230039

出　　处：《数学学报（中文版）》2000年第5期833-842,共10页Acta Mathematica Sinica：Chinese Series

基　　金：国家自然科学基金!(19670001)

摘　　要：为了克服在单参数双边截断型分布族中不能消除参数估计中在Ｂａｈａｒｄｕｒ意义下的超有效病态现象，本文提出了一种新的渐近效率．根据这种效率的定义，对一般 的参数函数，构造了适用的渐近中位无偏的渐近有效估计量．作为全文的理论基础，我们发现了渐近中位无偏估计的最优收敛速度．Extanding some Ibragimove-Hasminskii's idea, this paper points out that the phenomennon of superefficiency in Bahadur's sense can not be eliminated under one-parameter two-sided truncated distribution families. In order to eliminate this evil of superefficiency, a new kind of asymptotic efficiency have been established. Further the efficiencies of commonly used estimates have been calculated, and asymptotically median unbiased efficient estimates have been eonstructed.

关 键 词：超渐近有效性 渐近有效估计 单参数双边截断型分布族 参数估计 

分 类 号：O211.67[理学—概率论与数理统计] O212.1[理学—数学]