响应变量随机删失时函数型非参数分位数回归模型的估计

杨锦涛，凌能祥

(合肥工业大学数学学院, 安徽合肥 230601)

摘要

文章在响应变量随机删失时,研究了函数型非参数分位数回归模型,通过极小化逆概率加权分位数损失函数,构造模型中未知非参数函数的估计量。在一定的条件下,获得估计量的渐近正态性;通过模拟研究,验证了估计量的有效性。

关键词

函数型数据分析(FDA)；分位数回归；随机删失；逆概率加权；渐近正态

中图分类号：O212.7

文献标志码：A

文章编号：1003-5060（2023）05-0709-04

Estimation of nonparametric quantile regression model for functional data with censored response at random

YANG Jintao, LING Nengxiang

(School of Mathematics, Hefei University of Technology, Hefei 230601, China)

Abstract

In this paper, the nonparametric quantile regression model is presented to characterize the association between censored survival time and a set of functional predictors when response variables are censored at random, and estimates of nonparametric functions are obtained by minimizing the inverse probability weighted quantile loss function. Under some mild conditions, the asymptotic normality of the estimates is given. Simulation studies further verify the validity of the proposed model.

Keywords

functional data analysis (FDA); quantile regression; censoring at random; inverse probability weighting; asymptotic normal

收稿日期：2022-03-18

修回日期：2022-04-12

基金项目：国家自然科学基金资助项目（72071068）