Qi ZhengThursday, December 1, 2016 - 2:45pm
Statistics Department Colloquium
Where: LeConte College, Room 210
Speaker: Qi Zheng
Title: Censored quantile regression in high dimensional survival data
Censored quantile regression (CQR) has emerged as a useful regression tool for survival analysis. Stochastic integral based estimating equations are commonly used in the estimation of CQR, and pose new challenges in the analysis of CQR for high dimensional survival data. In this work, westudy the high dimensional CQR simultaneously over a continuum of quantile indices. We propose a two-step penalization procedure, which accommodates stochastic integral based estimating equations and properly addresses the associated complications. We establish the uniform convergencerates for the proposed estimators, and investigate the properties on weak convergence and variable selection. We conduct extensive numerical studies to confirm our theoretical findings and illustratethe practical utility of our proposals.Key Words: High dimensional data; Varying covariate effects; Model selection; Adaptive penalizedquantile regression.