Southwest Jiaotong University School of Mathematics

统计系

首页  >  学术科研  >  统计系  >  正文

新加坡国立大学张金廷教授和栗家量教授的学术报告

来源:   作者:统计系     日期:2021-11-09 08:38:46   点击数:  

报告时间:20211112日下午14

报告形式:线上报告(腾讯会议)

 

会议时间: 2021/11/12 13:30-16:30 (GMT+08:00)

点击专属链接入会,或添加至会议列表:

https://meeting.tencent.com/dw/Qb5PnC5MY4TA

会议 ID500 137 468

会议密码:211109



报告一 (北京时间20211112日下午14:00PM)

Title: Normal-Reference Tests for High-Dimensional Hypothesis Testing

Abstract: In the past two decades, much attention has been paid for high-dimensional hypothesis testing. Several centralized or non-centralized L2-norm based test statistics have been proposed.  Most of them imposed  strong assumptions on the underlying covariance structure of the   high-dimensional data so that the associated test statistics are asymptotically normally distributed.  In real data analysis, however, these assumptions are hardly checked so that the resulting tests have a size control problem when the required assumptions are not satisfied.  To overcome this difficulty,  in this talk,  we  investigate  a so-called  normal-reference test which can control the size well.  In the normal-reference test,  the null distribution of a test statistic is approximated with that of a chi-square-type mixture which is obtained from the test statistic when the null hypothesis holds and when the samples are normally distributed. The distribution of the chi-square-type mixture can be well approximated by a  three-cumulant matched χ2-approximation with the approximation parameters consistently estimated  from the data. Two simulation studies demonstrate that in terms of size control, the proposed normal- reference test performs well regardless of whether  the data are nearly uncorrelated, moderately correlated, or highly correlated and it performs much better than two  existing competitors. A real data example illustrates the proposed normal-reference test.


报告人简介:张金廷,新加坡国立大学概率统计系终身教授,博士生、博士后导师,华侨大学福建省闽江学者讲座教授。早年于北京大学取得学士学位,中国科学院应用数学所取得硕士学位,美国北卡罗来纳大学教堂山分校获得博士学位,美国哈佛大学博士后。先后在美国普林斯顿、罗彻斯特等大学做高级访问学者。张教授培养了数十名硕士和七位博士以及三位博士后,其主要学术成果发表在Annals of Statistics(世界统计年刊),JASA(美国统计学会杂志),JRSSB(英国皇家统计学会杂志),Statistics Sinica(统计学报)等统计学国际顶级期刊上,著有统计学专著《Analysis of Variance for Functional Data》和《Nonparametric  Regression Methods for Longitudinal Data Analysis》,以及一本学术论文集。现任和曾任多家学术期刊的编委,并多次担任大型国际会议的组织委员。张教授现在的研究领域包括非参数统计,纵向数据分析,函数数据分析,高维数据分析等。


报告二 (北京时间20211112日下午15:00PM)

Title: Multi-threshold Structural Equation Model

Abstract: In this paper, we consider the instrumental variable estimation for causal regression parameters with multiple unknown structural changes across subpopulations. We propose a multiple change point detection method to determine the number of thresholds and estimate the threshold locations in the two-stage least squares procedure. After identifying the estimated threshold locations, we use the Wald method to estimate the parameters of interest, i.e., the regression coefficients of the endogenous variable. Based on some technical assumptions, we carefully establish the consistency of estimated parameters and the asymptotic normality of causal coefficients. Simulation studies are included to examine the performance of the proposed method. Finally, our method is illustrated via an application of the Philippine farm households data for which some new findings are discovered.


报告人简介:栗家量,新加坡国立大学统计与数据科学系教授,同时担任杜克大学-新加坡国立大学医学院及新加坡眼科研究所的兼职教授。栗教授本科毕业于中科科学技术大学,博士毕业于美国威斯康大学。栗教授已发表科研论文160余篇, 主要学术成果发表在Annals of Statistics, JASA, Biometrics, Journal of Econometrics, JBES等统计学顶级杂志上。他最近的研究方向包括诊断医学,精准医学,非参数方法,统计学习与生存分析。他与合作者著有一本Chapman & Hall CRC Press2013年出版的专著Survival Analysis in Medicine and Genetics。 他的论文总引用量达到3600h-index 32。他在Biometrics Lifetime data Analysis等统计杂志担任过副主编。他曾获得新加坡国立大学的Young Scientist Award, 他也是美国统计学会(ASA)的会士(Fellow)和国际统计所(ISI)的选举会员(Elected member).