Asymptotic normality for a local composite quantile regression estimator of regression function with truncated data

被引:2
|
作者
Wang, Jiang-Feng [1 ]
Ma, Wei-Min [2 ]
Zhang, Hui-Zeng [3 ]
Wen, Li-Min [4 ]
机构
[1] Zhejiang Gongshang Univ, Dept Stat, Hangzhou 310036, Peoples R China
[2] Tongji Univ, Sch Econ & Management, Shanghai 200092, Peoples R China
[3] Hangzhou Normal Univ, Sch Sci, Hangzhou 310036, Zhejiang, Peoples R China
[4] Jiangxi Normal Univ, Dept Stat, Nanchang 330022, Jiangxi, Peoples R China
基金
中国国家自然科学基金;
关键词
Composite quantile regression; Local linear regression; Truncated data; Asymptotic normality; VARIABLE SELECTION; EFFICIENT;
D O I
10.1016/j.spl.2013.02.022
中图分类号
O21 [概率论与数理统计]; C8 [统计学];
学科分类号
020208 ; 070103 ; 0714 ;
摘要
In this paper, we construct a local linear composite quantile regression (CQR) estimator of regression function for left-truncated data, which extends the CQR method to the left-truncated model. The asymptotic normality of the proposed estimator is also established. The estimator is much more efficient than the local linear regression estimator for commonly-used non-normal error distributions via simulations. Crown Copyright (C) 2013 Published by Elsevier B.V. All rights reserved.
引用
收藏
页码:1571 / 1579
页数:9
相关论文
共 50 条