主  题：Variable selection procedure from multiple testing

内容简介：Summary. Variable selection plays an important role in statistical learning and scientific discoveries during the past ten years and multiple testing is a fundamental problem in statistical inference, also with wide applications in many scientific fields. This paper aims at figuring out a connection between adaptive lasso and multiple testing in the linear regression model, and aims at proposing a method based on the multiple testing to select variables and control the selecting error rate. Simulation studies show good performance of the proposed method on controlling the selecting error rate and achieving great powers under weak dependence.
Keywords: variable selection, multiple testing, adaptive lasso, false discovery

报告人：张宝学     教授    博导

                    2009年度教育部新世纪优秀人才支持计划

                    美国《数学评论》（“Mathematical Review”）特聘评论员

                    IMS-China执行理事

                    中国空间统计学会副理事长

时  间：2016-03-24    10:30

地  点：竞慧东楼305

举办单位：理学院