Statistics and Its Interface

Volume 15 (2022)

Number 3

Variable selection for time-varying effects based on interval-censored failure time data

Pages: 303 – 311

DOI: https://dx.doi.org/10.4310/21-SII687

Authors

Kaiyi Chen (Department of Statistics, University of Missouri, Columbia, Mo., U.S.A.)

Jianguo Sun (Department of Statistics, University of Missouri, Columbia, Mo., U.S.A.)

Abstract

Variable selection has recently attracted a great deal of attention and correspondingly, many methods have been proposed. In this paper, we discuss the topic when one faces interval-censored failure time data arising from a model with time-varying coefficients, for which there does not seem to exist a method. For the situation, in addition to identifying important variables or covariates, a desired feature of a variable selection method is to distinguish time-varying coefficients from time-independent ones, which also presents an additional challenge. To address these, a penalized maximum likelihood procedure is presented and in the proposed method, the adaptive group Lasso penalty function and B‑spline functions are used. The approach can simultaneously select between time-dependent and time-independent covariate effects. To implement the proposed procedure, an EM algorithm is developed, and a simulation study is conducted and suggests that the proposed method works well in practical situations. Finally it is applied a set of real data on Alzheimer’s disease that motivated this study.

Keywords

adaptive group LASSO, Cox model, interval-censored data

The full text of this article is unavailable through your IP address: 100.28.132.102

Received 4 April 2021

Accepted 16 June 2021

Published 14 February 2022