Quantile regression for functional partial linear models in high dimensions
报 告 人:: 朱仲义
报告地点:: 数学与统计学院415报告厅
报告时间:: 2017年07月02日星期日16:00-17:00

In this paper, we consider a functional partial linear quantile model in high dimensional scenario, where response is a scalar and predictors include multiple random processes and high-dimensional scalar covariates. A framework of regularization with two nonconvex penalty functions in the context of functional quantile regression are proposed formally, and the selection and estimation of important variables can be achieved by minimizing an double penalized functional quantile objective function. The approach proposed takes advantage of functional principal components which greatly facilitates implementation. We introduce a two-step technique for selecting tuning parameters, and establish the asymptotic oracle properties of the proposed estimators based on the difference convex analysis under some regularity conditions. The empirical performance and the usefulness of our approach are demonstrated through a large number of simulation studies and an application to air pollution data.

发 布 人:科研助理 发布时间: 2017-06-26
朱仲义,2015年获得教育部自然科学二等奖,2008-2010年两次访问美国北卡州立大学,2007年访问美国University of Illinois at Urbana Champaign统计学系,1999年10月2002.12,两次访问香港大学,发在Annals of The Institute of Statistical Mathematics等杂志发表论文80余篇。