报告时间：2018年5月30日（周三）14:30—15:00

报告地点：明德主楼1016会议室

报告题目：Sufficient Dimension Reduction with Nonignorable Nonresponse

报告人简介：Jun Shao  University of Wisconsin-Madison, USA

邵军教授 1987 年 8 月获美国威斯康星 - 麦迪逊分校统计学博士学位，1996 年获美国数理统计学会 Fellow，1999 年获美国统计学会 Fellow，多次获得美国自然科学基金，曾任 JASA、Statistica Sinica 副主编，Journal of Multivariate Analysis 和 Sankhya 联合主编，现任 Journal of Nonparametric Statistics 主编，Journal of System Science and Complexity 联合主编，2017 年联合创立 Statistical Theory and Related Fields 并担任总编辑。曾担任美国威斯康星 - 麦迪逊分校统计系系主任（2005-2009）、泛华统计学会会长（2007），现兼任美国国家统计局高级研究员，并任美国多家制药厂的统计顾问，现为美国威斯康星 - 麦迪逊分校统计系教授。邵教授的 6 本统计学专著和课本之一的《数理统计》已成为数理统计理论名著，并成为北美和中国多个大学的统计学研究生教材。自 1987 年以来邵教授共发表学术论文 180 余篇，其中 50 余篇为医药统计方面论文，在重抽样技术、变量选择、生物统计和缺失数据的统计处理等方面做了大量的开创性工作。

报告摘要：We consider sufficient dimension reduction in the situation where a response variable has nonignorable nonresponse, the joint distribution of the response variable and the associated co- variate vector is unspecified, and the nonresponse propensity is semi-parametric. To handle the nonignorable nonresponse, we propose a three-step procedure to sufficiently reduce the covariate dimension, based on some identities obtained under the semi-parametric propensity. After covariate dimension reduction, we further derive nonparametric estimators of unknown parameters in the unspecified distribution of the response variable and covariates. A key part of this derivation is utilizing the results from sufficient dimension reduction to find a nonresponse instrument, a linear function of covariates that is related to the response variable but can be excluded from the propensity. Asymptotic normality of the proposed estimators are established. We evaluate the performance of the proposed estimators in a Monte Carlo study and illustrate our method in an application to AIDS Clinical Trials Group Protocol 175 data.