报告一

题目：国家社科基金申报心得分享

报告人： 杨青龙

时间：2020年10月16日 8：30-9：30

地点：腾讯会议 ID 748307864

摘要：主要从国家社科基金的评审流程、社科基金选题以及项目论证等几部分分享自己申请人文社科类项目的心得。

杨青龙简介：

中南财经政法大学统计与数学学院副院长，数理与金融统计学系副教授，校“文澜青年学者”。2011年6月在武汉大学数学与统计学院取得博士学位，2011年9月-2012年9月在美国爱荷华大学统计与精算系从事博士后研究。2012年9月开始在中南财经政法大学统计与数学学院任教。主要从事生存分析，高维数据分析，机器学习和统计计算等方向的研究。主持国家自然科学基金、国家社科基金、教育部人文社科一般项目、国家统计局统计科学研究重点项目、中央高校业务经费、湖北省农业普查项目和湖北省金融统计学会重点课题等10余项课题。在《IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS》 《SCIENCE CHINA Mathematics》 《数量经济技术经济研究》等期刊发表论文近20篇。

报告二

标题：Describing and analyzing the multivariate count data without 0 observations: A new multivariate ZTPLN model

报告人： 刘寅

时间：2020年10月16日 9：30-10：30

地点：腾讯会议 ID 748307864

摘要：It is well known that the multivariate Poisson log-normal (PLN) distribution (Aitchison & Ho 1989) has a wide range of applications in modeling multivariate count data due to its flexible correlation structure. However, in many practical problems, we often encounter multivariate count data without zero vectors. In this paper, for the first time, we propose a new multivariate zero-truncated PLN distribution to model multivariate count data without zero vectors. An expectation–maximization algorithm based on the method of stochastic representation is developed to derive the maximum likelihood estimates of parameters of interest. Third, testing hypotheses on an inner correlation, on a degenerated form and on an isotropic form are provided. Fourth, a new DA-SIR algorithm is developed for Bayesian posterior sampling. In addition, we also investigate the multivariate zero-truncated PLN regression model by using the Bayesian method. Finally, a data set of two-dimensional fatal train accidents in Britain is employed to demonstrate the proposed methods.

刘寅简介： 中南财经政法大学统计与数学学院，讲师，硕士生导师。中国现场统计研究会试验设计分会理事、中国数学学会均匀设计分会理事。研究领域：敏感数据调查分析、多元统计分析。在《Journal of the Computational and Applied Mathematics》《Computational Statistics and Data Analysis》《Statistical Methods in Medical Research》《Biometrical Journal》《Statistical Papers》等国际权威期刊上发表论文十余篇。主持和参与多项国家自然科学基金项目。

报告三

题目: Doubly Divided Massive Data for Prediction Using Model Aggregation

报告人： 吴远山

时间：2020年10月16日 10：30-11：30

地点：腾讯会议 ID 748307864

摘要: Nowadays, massive data are often featured with high dimensionality as well as huge sample size, which typically cannot be stored in a single machine and thus make both analysis and prediction challenging. We propose a distributed gridding model aggregation (DGMA) approach to predicting the conditional mean of response, which overcomes the storage limitation of a single machine and the curse of high dimensionality. Specifically, on each local machine that stores partial data of relatively moderate sample size, we develop the model aggregation approach by splitting predictors wherein a greedy algorithm is developed. To obtain the optimal weights across all local machines, we further design a distributed and communication-efficient algorithm. Our procedure effectively distributes the workload and dramatically reduces the communication cost. Theoretically, we establish the prediction error bound of the DGMA method, which can be explicitly expressed in terms of local sample size and communication rounds. We further show that if local sample size or communication rounds are sufficiently large, the proposed method can reach the prediction error bound of the oracle global method that has access to full data. Extensive numerical experiments are carried out on both simulated and real datasets to demonstrate the feasibility of the DGMA method. This is a joint work with Baihua He, Yanyan Liu and Guosheng Yin.

吴远山简介：中南财经政法大学统计与数学学院教授、博士生导师。吴远山教授博士毕业于武汉大学，研究方向为高维数据分析、生存分析。主持多项国家级和省部级科研项目。现已在统计方向顶级期刊Journal of the American Statistical Association, Biometrika, Biometrics等发表学术论文20余篇。