学术动态

Variable selection for high dimensional nonlinear models based on weighted composite quantile regression

报 告 人:: 蒋学军
报告地点:: 数学与统计学院四楼学术报告厅
报告时间:: 2017年04月14日星期五9:30-10:30

报告简介:

   Quantile regression is a method of natural regression analysis which uses the central trend and the degree of statistical distribution to obtain a more comprehensive and powerful analysis. In this talk, we propose a weighted composite quantile regression (WCQR) estimation approach and study model selection for high dimensional nonlinear models. The WCQR is augmented using a data-driven weighting scheme. With the error distribution unspecified, the proposed estimators share robustness from quantile regression and achieve nearly the same efficiency as the oracle maximum likelihood estimator for a variety of error distributions including the normal, mixed-normal, Student’s t, Cauchy distributions and etc. Based on the proposed WCQR, we use the adaptive-LASSO and SCAD regularization to simultaneously estimate parameters and select models. Under regularity conditions, we establish asymptotic equivalency of the two model selection methods and show that they perform as well as if the correct sub-models are known in advance.

   We also suggest an updated interior algorithm for fast implementation of the proposed methodology. Simulations are conducted to compare different estimators, and a real example is used to illustrate their performance.

主讲人简介:

蒋学军,南方科技大学数学系助理教授&博士生导师,深圳市海外高层次人才“孔雀计划”入选者,2009年博士毕业于香港中文大学统计学系;2010-2011在香港中文大学从事博士后研究,2011年10月至2013年6月历任中南财经政法大学数理与金融统计系讲师,副教授&教研室主任,研究生导师,EMBA指导教师。2013年7月进入南方科技大学,获得国家自然科学基金项目,广东省自然科学基金项目及深圳市技术委托开发项目资助各一项,现任广东省金融教指委委员 。 蒋学军博士主要从事高维数据分析,非参数与半参数模型,金融计量经济学、贝叶斯应用等研究工作。在Statistica Sinica, Canadian Journal of Statistics,Statistical Methods in Medical Research, Bayesian Analysis, Econometrics Journal, Insurance:Mathemathics& Econmics 等统计学著名刊物上发表SCI&SSCI期刊论文近30篇。

专题网站Project site