Statistics and Its Interface

Volume 9 (2016)

Number 2

Transformed linear quantile regression with censored survival data

Pages: 131 – 139

DOI: https://dx.doi.org/10.4310/SII.2016.v9.n2.a1

Authors

Rui Miao (Institute of Applied Mathematics, Academy of Mathematics and Systems Science, Chinese Academy of Sciences, Beijing, China)

Liuquan Sun (Institute of Applied Mathematics, Academy of Mathematics and Systems Science, Chinese Academy of Sciences, Beijing, China)

Guo-Liang Tian (Department of Statistics and Actuarial Science, University of Hong Kong)

Abstract

Quantile regression provides a flexible method for analyzing survival data, and attracts considerable interest in survival analysis. In this article, we propose a new inference procedure for a class of power-transformed linear quantile regression models with survival data subject to conditionally independent censoring, and present a two-stage algorithm that is computationally simple and easy to implement. Consistency and asymptotic normality of the resulting estimators are established, and a simple resampling-based inference procedure is developed for variance estimation. The finite-sample behavior of the proposed methods is examined through extensive simulation studies. An application to a real data example from a health maintenance organization is provided.

Keywords

censored survival data, Box-Cox transformation, martingale, quantile regression, resampling

2010 Mathematics Subject Classification

Primary 62N01, 62N02. Secondary 62G20.

Published 4 November 2015