主 题：Marginal Empirical Likelihood and Independence Feature Screening
主讲人：常晋源博士
主持人：马昀蓓博士
时 间：2013年12月26日下午14：00-15：00
地 点：通博楼B座212学术会议室
主办单位：统计学院 统计研究中心 科研处
主讲人简介：
常晋源，墨尔本大学数学与统计系Research fellow，合作导师为Peter Hall教授。2013年7月毕业于北京大学光华管理学院商务统计与计量经济系，师从陈松蹊教授。2012年7月获国际数理统计协会（Institute of Mathematical Statistics）颁发的Laha Award，2013年10月获中国数学会颁发的第十一届钟家庆数学奖。
内容提要：
In this talk, I will mainly focus on how to construct a unified screening procedure via empirical likelihood for general parametric models that defined by general moment conditions. Such model setting includes linear models and generalized linear models as special cases. Different from most existing feature screening approaches that rely on the magnitudes of some marginal estimators to identify true signals, the proposed screening approach is capable of further incorporating the level of uncertainties of such estimators. Such a merit inherits the self-studentization property of the empirical likelihood approach, and extends the insights of existing feature screening methods. The theoretical results and extensive numerical examples by simulations and data analysis demonstrate the merits of the marginal empirical likelihood approach. If time is enough, I will show how to extend this idea to do feature screening in general nonparametric and semi-parametric models. This talk is based on my two papers with Cheng Yong Tang and Yichao Wu.
