学术活动
当前位置: 首页 >> 新闻中心 >> 学术活动 >> 正文

Multi-threshhold accelerated failure time model

2018/05/08 09:54  点击:[]


报告人: 栗家量 (National University of Singapore)

 

 : 2018年5月14日   15:00---16:00

 

 : 理科楼 LD202

 

 :A two-stage procedure for simultaneously detecting multiple thresholds and achieving model selection in the segmented accelerated failure time (AFT) model is developed in this paper. In the first stage, we formulate the threshold problem as a group model selection problem so that a concave 2-norm group selection method can be applied. In the second stage, the thresholds are finalized via a refining method. We establish the strong consistency of the threshold estimates and regression coefficient estimates under some mild technical conditions. The proposed procedure performs satisfactorily in our simulation studies. Its real world applicability is demonstrated via analyzing a follicular lymphoma data.

 

报告人简介: 栗家量教授就职于新加坡国立大学,主要研究领域为非参数统计、函数型数据分析、高维统计推断等,已发表包含Journal of the Royal Statistical Society Series B, Annals of Statistics, Journal of the American Statistical Association等顶级期刊在内的论文60余篇,目前担任Biometrics, Lifetime Data Analysis等国际权威期刊的Associate Editor。

 

学院联系人: 徐建文

 

欢迎广大师生积极参与!

上一条:Einstein Finsler Metric... 下一条:非线性偏微分方程小型研讨会

关闭

数学与统计学院 College of Mathematics and Statistics
(C)Copyright: College of Mathematics and Statistics All rights reserved
地址:重庆大学虎溪校区理科大楼D栋 邮编:401331