CFE 2019: Start Registration
View Submission - CMStatistics
Title: A quantile localized approach to the accelerated failure time model on survival data with time-dependent covariates Authors:  Tony Sit - The Chinese University of Hong Kong (Hong Kong) [presenting]
Abstract: The purpose is to discuss a generalization of the accelerated failure time model for survival data subject to right censoring, which is independent of the actual lifetime conditional on possibly time-varying covariates. We require the homogeneous conditional quantile assumption on the lifetime for a localized range of quantile levels, instead of assuming it hold globally. By introducing a class of weighted rank-based estimation procedure, the framework allows a quantile localized inference on the covariate effect with less stringent assumption. Meanwhile, the form of the proposed estimating equations can be viewed as a generalization of its counterpart under the accelerated failure time model with time-varying covariates. Numerical studies demonstrate that the proposed estimator overperforms current alternatives under various settings in terms of smaller empirical bias and standard deviation. A perturbation-based resampling method is also provided to reconcile the asymptotic distribution of the parameter estimates. Finally, consistency and weak convergence of the proposed estimator is established via empirical process theory.