双侧截断数据的中位数回归(英文)

在很多实际应用中,个体寿命时间可能被同时左截断与右截断.本文在左右截断变量都能被观察到的假设下,提出了一种半参数推断方法,来分析协变量对于相应寿命时间或其某种变换的中位数的影响,并讨论了所得估计量的渐近性质.此外,本文还提供了一种基于经验似然的回归参数推断方法,并讨论了将这些方法推广到经典双侧截断数据的可能性.一些模拟计算被用于展示这些方法的有效性.

Median Regression for Left and Right Censored Data

In many applicationsinvolving follow-up studies, individuals' lifetimes may be subjectto both left and right censoring. In this paper we considerregressing the median of the lifetime variable or a transformationthereof on the corresponding covariates when the lifetimes aresubject to both left and right censoring and the values of bothcensoring variables are always observable. A semiparametricinferential procedure is proposed and the asymptotic properties ofthe proposed method are discussed. We also propose an alternativeempirical likelihood based inferential procedure for the regressioncoefficients vector. Moreover, we make some discussion on classicaldoubly censored data whose censoring variables cannot always beobserved. The proposed methods are illustrated by some simulationstudies.