A1261
Title: Quantile regression with asynchronous longitudinal data
Authors: Xuerui Li - Beijing Normal University (China) [presenting]
Yanyan Liu - Wuhan University (China)
Yuanshan Wu - Zhongnan University of Economics and Law (China)
Lixing Zhu - Beijing Normal University (China)
Abstract: In many biomedical longitudinal studies, time-dependent responses and covariates are observed asynchronously within subjects. And the biomedical data often presents heteroscedasticity with outliers and a skewed distribution in response. Due to the fact that quantile regression is generally robust in handling skewed responses in heteroscedastic data and flexible to accommodate covariate-response relationships, we consider quantile regression modelling to include time-invariant and time-varying coefficients for such longitudinal data. Asymptotic properties are established, including consistency and weak convergence. Simulations studies suggest the good finite-sample performance of the proposed method. The practical example concludes more comprehensive results under the proposed estimation when comparing directly using synchronous data analysis methods and mean regression.
