报告人: 蒋学军副教授 南方科技大学
时间：2023年11月17日 10:30-12:00
地点：数学与信息学院201报告厅
Abstract: Strong correlation among predictors poses a great challenge in the analysis of ultra-high dimensional data. This leads to an increase in the computation time for screening active variables and a decrease in selection accuracy. To address this issue, we propose an innovative two-stage feature selection procedure and its derivative versions based on sparsity assumption. This approach initially screens important features by ranking quantile ridge estimation and subsequently employs a likelihood-based post-screening selection strategy to refine variable selection. This procedure is simple to implement and maintains the robustness of quantile regression. In addition, it possesses other numerous other desirable properties from theoretical and computational standpoints. Theoretically, we establish the strong consistency of feature selection for the proposed methods under some regularity conditions. In empirical studies, we assess the finite sample performance of our methods by comparing them with marginal quantile utility screening approaches and existing penalized quantile regression methods. Furthermore, we apply our methods to identify genes associated with anticancer drug sensitivities.
报告人简介: 蒋学军，南方科技大学统计与数据科学系副教授（长聘），博士生导师，于2009年博士毕业于香港中文大学统计系，2009-2010在港中文从事博士后研究，2010-2013任中南财经政法大学副教授，于2013年07月加入南方科技大学，入选深圳市海外高层次人才孔雀计划（2016），曾获南方科技大学杰出教学奖（2018），深圳市优秀教师（2018），主持和完成国家（广东省）自然科学基金、深圳市基础研究面上项目等10余项。其主要研究方向包括分位数回归、变量选择、假设检验、高维统计推断，金融统计与计量等，已发表包括Bernoulli Journal , Statistica Sinica, Econometrics Journal, Science China-Mathematics等在内的SCI&SSCI论文50余篇，授权专利1项，并出版英文教材一部。国内学会任职主要有中国现场统计研究会-教育统计与管理分会副理事长，多元分析应用专业委员会秘书长等。
欢迎广大师生参加！
