统计学系系列讲座之235-237期

 

统计学系系列讲座之236

 

时间:2016年10月21日(星期五)上午10:15-11:00

地点:史带楼204室

主持人:朱仲义 教授 复旦大学管理学院统计学系

:Estimating large covariance and precision matrices from temporally dependent data

主讲人:Prof. Bin Nan 

University of Michigan

主讲人简介:Bin Nan教授是密歇根大学生物统计系和统计系的教授。研究领域主要涉及半参数统计,纵向数据,高维数据,缺失数据,两阶段抽样设计,在Journal of the American Statistical Association, Biometrika,Biometrics,Statistica Sinica等杂志已发表100余篇论文。Bin Nan教授曾担任Statistica Sinica杂志副主编,现为Statistics in Biosciences杂志和The Canadian Journal of Statistics杂志副主编。 

摘要:We consider the estimation of large covariance and precision matrices from high-dimensional sub-Gaussian observations with slowly decaying temporal dependence that is bounded by certain polynomial decay rate. The temporal dependence is allowed to be long-range so with longer memory than those considered in the current literature. We show that several commonly used methods for independent observations can be applied to the temporally dependent data. The rates of convergence are obtained, and the properties of sparsistency and sign-consistency are also established. A gap-block cross-validation method via low-dimensional submatrices is proposed for the tuning parameter selection, which performs well in simulations.

 

统计学系系列讲座之237

 

时间:2016年10月20日(星期四)下午16:00-17:00

地点:史带楼502室

主持人:张新生 教授 复旦大学管理学院统计学系

:Nonparametric Shrinkage Estimation

主讲人:Professor Cun-Hui Zhang(张存惠)

Department of Statistics and Biostatistics, Rutgers University

主讲人简介:张存惠教授是Rutgers University统计和生物统计系的Distinguished Professor,他是IMS和ASA的Fellow。其主要研究方向为: 高维数据 (High dimensional data)、经验Bayes (Empirical Bayes)、半参数与非参数方法(Semiparametric and Nonparametric Methods)、生存分析(Survival Analysis)、网络数据(Network Data)等。

摘要:We revisit the classical problem of estimating the mean vector of three or more observations. James and Stein (1961) claimed that given a known upper bound for the fourth moment of the noise, a shrinkage estimator of Stein (1956) always has a smaller total mean squared error than the observed vector itself as a naive estimator, provided that certain parameters of the shrinkage factor are properly specfied. James and Stein (1961) further commented that “It would be desirable to obtain explicit formulas for estimators one can seriously recommend”  in this setting. In the present paper, we provide some explicit solutions to this nonparametric shrinkage estimation problem.

 

统计学系

2016-10-18