三种两样本经验Euclidean似然方法及其比较  

Three Methods of Empirical Euclidean Likelihood for Two Sample and Their Comparison

作　　者：林路[1] 张润楚[1] 

机构地区：[1]南开大学数学科学学院统计学系,天津300071

出　　处：《应用概率统计》2002年第4期393-399,共7页Chinese Journal of Applied Probability and Statistics

基　　金：本文得到国家自然科学基金(编号10171051);南开大学青年教师基金(编号12902516)的资助

摘　　要：本文研究三种两样本经验Euclidean似然方法:基于两个无偏估计函数的经验Euclidean似然方法;基于一个无偏估计函数和一个历史经验估计的经验Euclidean惩罚似然方法;基于两个经验估计的加权和方法.我们研究了这些方法的强相合性,渐近正态性和渐近有效性.研究表明,这三种方法是同等渐近有效的.This paper studies the three methods of empirical Euclidean likelihood for two sample: empirical Euclidean likelihood based on two unbiased estimating functions; empirical Euclidean penalized likelihood based on an unbiased estimating function and an empirical estimator; weighted sum based on two empirical estimators. The strong consistency, asymptotic normality and asymptotic efficiency of these methods are proved. The results show that asymptotic efficiency of each of the three methods is the same.

关 键 词：经验Euchlidan似然 经验Euclidean惩罚似然 强相合性 渐近正态性 渐近有效性 

分 类 号：O212.1[理学—概率论与数理统计]