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.
出处
《数学学报（中文版）》
SCIE
CSCD
北大核心
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)
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
One-parameter two-sided truncated families
Superefficiency
Efficient estimate
Asymptotically median unbiased estimate
