讲座题目：Permutationtest under rotating sampling plan with clustered data
主讲人：Professor Jiahua Chen 英属哥伦比亚大学
讲座时间：2020年8月4日（周二）上午10：00―12：00
参与方式：腾讯会议会议 ID：194 559 343；会议直播网址: https://meeting.tencent.com/l/yDaJnJZA5UiC
主讲人简介：
Professor Jiahua Chen iscurrently a Canada Research Chair, Tier I, in statistical inference and aProfessor in the Department of Statistics, University of British Columbia, Vancouver, Canada. Dr. Jiahua Chenreceived his Master's degree from the Institute of System's Science in Jan1985, and Ph./ D degree from the University of Wisconsin-Madison in July 1990under the supervision of Professor Jeff Wu.
His research interestsinclude Finite Mixture Models, Statistical Genetics, Empirical Likelihood, SurveyMethodology, Design of Experiment, and others.
He has published hisresearch in most major statistical journals such as The Annals of Statistics,The Journal of American Statistical Association, The Journal of the RoyalStatistical Society, Biometrika and Technometrics.
He served as thepresident of the International Chinese Statistical Association in 2005, and thepresident of the Section of Survey Methodology, the Statistical Society ofCanada, 2007.
He was a member ofthe Grant Selection Committee of the Natural Science and Engineering ResearchCouncil of Canada.
He is an electedfellow of the Institute of Mathematical Statistics and the American StatisticalAssociation.
He is the recipientof the CRM-SSC Prize in Statistics for his outstanding contributions to thestatistical sciences.
This prestigious award, jointly sponsored bythe Statistical Society of Canada(SSC) and the Centre de recherchesmathématiques de Montréal (CRM), is given each year to a Canadian statisticianin recognition of outstanding contributions to the discipline during therecipient's first 15 years after earning a doctorate.
He received the Gold Medal, the top prize ofthe Statistical Society of Canada in 2014, the International ChineseStatistical Association distinguished achievement award in 2016.
讲座摘要：
Consider a population madeof clusters of sampling units,evolving temporally, spatially or according toother dynamics. We wish to monitor the evolution of its means, variances,mediansor other population parameters.
In a real world, clustered data are oftencollected via a rotating plan for administrative convenience andinformativeness. Under rotating plans, not only the observations in the sameclusters are correlated, but also observations on the same unit collected indifferent occasions {/blue are correlated}.
Ignoring the correlation structure in theclustered data can lead to invalid inference procedures. Accommodating suchcluster structures in parametric models is difficult or can lead tounacceptable level of misspecification risk.
In this paper, we explore the exchangeabilityin clustered data collected via rotating sampling plan to develop a permutationscheme for testing various hypotheses of interest.
We also introduce a semi-parametric densityratio model to facilitate the multiple population structure in rotatingsampling plans.
The combination ensures the validity of theinference methods while {/blue extracting} maximum information carried by therotating sampling plan. Simulation study indicates that the proposed testsfirmly control the type I error when data are clustered (or not).
The use of the density ratio model improvesthe power of the tests. We also investigate the influence of the choice of basisfunction in the density ratio modeland conclude that theefficiency gain is widely observed.
