讲座题目： How to make model-free feature screening approaches for full data applicable to the case of missing response？

讲座专家：王启华，中国科学院数学与系统科学研究院研究员

讲座时间：2017年3月16日 16:00-17:30

讲座地点：对外经济贸易大学科研楼配楼五层会议室

讲座内容简介：

It is quite challenge to develop model-free feature screening approaches {/it directly} for missing response problems since the existing standard missing data analysis methods cannot be applied directly to high dimensional case. This paper develops some novel methods by borrowing information of missingness indicators such that any  feature screening procedures for ultrahigh-dimensional covariates with full data can be applied to missing response case.

The first method is the so-called missing indicator imputation screening which is  developed by proving that the set of the active predictors of interest for the response  is a subset of the active predictors for the product of the response and missingness indicator under some mild conditions. As an alternative, another  method, which is called as Venn diagram based approach, is also developed for obtaining a feature screening estimator of the active predictor set of interest. The sure screening property is proven for both the methods.

It is shown that the complete case (CC) approach can also keep the sure screening property of any feature screening approach with sure screening property.  A simulation study was conducted to compare the proposed methods with the CC approach and indicates that the proposed missing indicator imputation feature screening  method outperforms the CC method and the Venn diagram based method in some significant cases and is quite competitive in some other cases.

讲座专家介绍：

王启华，中国科学院数学与系统科学研究院研究员，博士生导师，国家杰出青年基金获得者，教育部 “长江学者奖励计划”特聘教授，中科院“百人计划”入选者， 国际统计研究会当选会员（Elected Member of ISI).