Position : homepage  > 2019

Statistics |Robust post inference of high dimensional linear models

Speaker :

Time :

Location :

题目:Robust post inference of high dimensional linear models

报告人:林媛媛 (香港中文大学)

时间:2019.10.28(周一)上午9:00

地点:紫金港校区管理学院行政楼14楼1417报告厅

摘要:

Robust inference is of key importance to large scale inference.  Under the high dimensional linear regression with an intercept term, we propose a robust post-selection inference method based on the Huber loss  for the regression parameters,  when the error distribution is heavy-tailed and asymmetric  that is common in various scientific fields.

The asymptotic properties of the resulting estimator are established under mild conditions. Statistical tests for low-dimensional parameters or individual coefficient in the high dimensional linear model are also studied. Simulation studies demonstrate desirable properties of the proposed method. An application to a genomic dataset about riboflavin  production rate is provided.

欢迎广大师生踊跃参加!

联系人:蒋杭进(jianghj@zju.edu.cn

                   浙江大学数据科学研究中心 

报告人简介:

林媛媛博士毕业于香港科技大学数学系,现为香港中文大学统计系助理教授. 她的主要研究方向为高维数据分析,生存分析等,目前已经在JASA, Biometrika,Statistica Sinica等国际知名期刊上发表论文17篇。

Date: 2019-10-22 Visitcount : 474