• 统计研究中心
当前位置: 首页> 系列讲座> 正文

上海交通大学罗珊副教授:Feature Selection by Canonical Correlation Search in High-Dimensional Multiresponse Models With Complex Group Structures

光华讲坛——社会名流与企业家论坛第 5613 期

线上讲座

Feature Selection by Canonical Correlation Search in High-Dimensional Multiresponse Models With Complex Group Structures

主讲人上海交通大学罗珊副教授

主持人统计学院林华珍教授

时间2020年12月29日(周二)上午10:30-11:30

直播平台及会议ID腾讯会议,146 946 608

主办单位:统计研究中心和统计学院 科研处

主讲人简介:

罗珊,新加坡国立大学统计学博士,密歇根大学生物统计系访问学者。现为上海交通大学数学科学学院长聘教轨副教授。主要研究领域为高维向量和矩阵数据、函数型数据、时空数据中的分类问题、模型选择标准和变量选择方法。文章主要发表在Journal of the American Statistical Association,Statistica Sinica,Journal of Multivariate Analysis,Annals of the Institute of Statistical Mathematics,Computational Statistics and Data Analysis, Journal of Statistical Planning and Inference等期刊上。详情请见其个人主页:http://www.math.sjtu.edu.cn/faculty/show.php?id=136


内容提要:

   High-dimensional multiresponse models with complex group structures in both the response variables and the covariates arise from current researches in important fields such as genetics and medicine. In this talk, I will introduce a novel approach named the sequential canonical correlation search (SCCS) procedure. In the SCCS procedure, the nonzero group by group blocks of regression coefficients are searched stepwise using a canonical correlation measure. Each step of the procedure consists of a block selection and a sparsity identification. The model selection criterion, EBIC, is used as the stopping rule of the procedure. We establish the selection consistency of the SCCS procedure and conduct simulation studies for the comparison of existing methods. A real example in genetic studies is also considered.

响应变量和协变量具有复杂组结构的高维多响应模型是当前遗传学、医学等重要领域研究的热点。本次报告将介绍一种新的方法,称为序贯点则相关搜索(SCCS)方法。SCCS方法使用点则相关测度逐步搜索回归系数的非零组。该方法的每一步包括块选择和稀疏性识别,最后采用模型选择准则EBIC作为方法的停止规则。我们建立了SCCS方法的选择相合性,并进行了模拟研究和遗传学中的实际数据分析,然后与现有的方法进行了比较。


上一条:南京审计大学吕绍高教授:Nonparametric optimality for large compressible deep neural networks under quadratic loss functions

下一条:微众银行人工智能部副总经理吴海山博士:人工智能和另类数据在资产管理中的应用