TY - JOUR
T1 - Variable Selection for Case-cohort Studies with Informatively Interval-censored Outcomes
AU - Du, Mingyue
AU - Zhao, Xingqiu
AU - Sun, Jianguo
N1 - Funding Information:
The authors wish to thank the Co-Editor Dr. Byeong Park, the Associate Editor and two reviewers for their many helpful and insightful comments and suggestions that greatly improved the paper. They also want to thank Dr. Peter Gibert for providing the HVTN 505 trial data. This research was supported in part by the National Natural Science Foundation of China ( 12101522 ), the Research Grants Council of Hong Kong ( 15303319 ), the CAS AMSS-PolyU Joint Laboratory of Applied Mathematics , and the Hong Kong Polytechnic University ( P0038663 ).
Publisher Copyright:
© 2022 Elsevier B.V.
PY - 2022/8
Y1 - 2022/8
N2 - Variable selection has recently attracted a great deal of attention and in particular, a couple of methods have been proposed for general interval-censored failure time data or the interval-censored data arising from case-cohort studies. However, all of them have some limitations or apply only to limited situations. Corresponding to these, a new, more general variable selection approach is proposed under a class of flexible semiparametric transformation models that allows for dependent interval censoring. In particular, the oracle property of the method under the broken adaptive ridge penalty function is established and for its implementation, a novel EM algorithm is developed. Also a simulation study is performed and suggests that the proposed approach works well in practical situations. Finally the method is applied to a HIV trial that motivated this study.
AB - Variable selection has recently attracted a great deal of attention and in particular, a couple of methods have been proposed for general interval-censored failure time data or the interval-censored data arising from case-cohort studies. However, all of them have some limitations or apply only to limited situations. Corresponding to these, a new, more general variable selection approach is proposed under a class of flexible semiparametric transformation models that allows for dependent interval censoring. In particular, the oracle property of the method under the broken adaptive ridge penalty function is established and for its implementation, a novel EM algorithm is developed. Also a simulation study is performed and suggests that the proposed approach works well in practical situations. Finally the method is applied to a HIV trial that motivated this study.
KW - Case-cohort studies
KW - Informative censoring
KW - Semiparametric transformation model
KW - Variable selection
UR - http://www.scopus.com/inward/record.url?scp=85127526484&partnerID=8YFLogxK
U2 - 10.1016/j.csda.2022.107484
DO - 10.1016/j.csda.2022.107484
M3 - Journal article
AN - SCOPUS:85127526484
SN - 0167-9473
VL - 172
SP - 1
EP - 17
JO - Computational Statistics and Data Analysis
JF - Computational Statistics and Data Analysis
M1 - 107484
ER -
