A0912
Title: Censored quantile regression with time-dependent covariates
Authors: Chi Wing Chu - City University of Hong Kong (Hong Kong)
Tony Sit - The Chinese University of Hong Kong (Hong Kong)
Zhiliang Ying - Columbia University (United States)
Chi Wing Chu - City University of Hong Kong (Hong Kong) [presenting]
Abstract: A new class of censored quantile regression models is proposed with time-dependent covariates for right-censored failure time data. While time-dependent covariates naturally arise in time-to-event analysis, existing works in the literature discuss treatments for data collected either under an independent censoring mechanism or a longitudinal setting. The formulation extends the current scope so that the conventional setting of time-dependent covariates can be properly handled. The new framework also generalizes the definition of quantiles and offers a new dynamic perspective for interpretation. Asymptotic properties of the recursive estimator are established. Numerical studies are also presented to illustrate the effectiveness of the proposal.
