Asymptotic Properties of a Modified Semi-Parametric MLE in Linear Regression with Right-Censored Data

被引：0
|
作者：
WONG George Y.C. [1 ]
机构：
[1] Strang Cancer Prevention Center.428 E 72nd Street.New York.NY 10021.USA
来源：
ActaMathematicaSinica(EnglishSeries) | 2002年 / 18卷 / 02期
关键词：
Consistency; Super-efficiency; Semi-parametric MLE;
D O I：
暂无
中图分类号：
O212.1 [一般数理统计];
学科分类号：
020208 ; 070103 ; 0714 ;
摘要：
Consider a linear regression model Y=β’X+e.where Y may be right censored and thecdf F+o of e is unknown.We show that a modified semi-parametric MLE.denoted by .is stronglyconsistent under certain regularity conditions.Moreover.if Fis discontinuous,then P(≠βi.o.)=0,which means that P(=βif the sample size is large)=1.The latter property has not been reportedfor the existing estimators.By contrast,most estimators,such as the Buckley-James estimator andM-estimators .satisfy that P(≠βi.o.)=1.
引用
收藏
页码：405 / 416
页数：12
相关论文
共 50 条
12345