统计学系系列讲座之288-291期

 

统计学系系列讲座之288期

 

时间:2018年3月27日(周二)15:30-16:30

地点:史带楼410室

主持人:夏寅 青年研究员 复旦大学管理学院统计学系

主题:Simultaneous variable selection and class fusion with penalized distance criterion based classifiers

主讲人:王启华 教授 中国科学院

简介:王启华,中国科学院核心骨干特聘研究员,博士生导师,国家杰出青年基金获得者,教育部**学者奖励计划特聘教授,中科院“百人计划”入选者,国际统计研究会当选会员(elected member)。主要从事生存分析、缺失数据分析、高维数据统计分析及非-半参数统计推断等方面的研究。出版专著两部,发表论文百余篇,其中90多篇发表在 The Annals of Statistics,  JASA及Biometrika等国际重要刊物, 2014、2015、2016与2017连续4年被Elsevier列入中国高被引学者榜单, 是一些国际与国内刊物的主编与编委。

摘要:In this paper, we propose two new methods to solve the problem of constructing sparse multiclass classifiers and determining corresponding discriminative variables for each pair of classes simultaneously in the high-dimensional setting. In contrast to many existing multiclass classifiers, which can only select informative variables for classification, we can understand roles of the selected variables in separating particular pairs of classes more profoundly by using different penalties. Different from  Guo (2010, Biostatistics) and Xu et al. (2015, Biometrika), which are based on the separate estimation of the precision matrix and mean vectors, we propose to construct classifiers by estimating products of the precision matrix and mean vectors or all discriminant directions directly with more appropriate penalties. This leads to the use of the distance criterion instead of the log-likelihood used in existing literature. With the proposed methods, we can not only consistently select informative variables for classification but also consistently identify corresponding discriminative variables for each pair of classes.  More importantly, our methods attain asymptotically the optimal misclassification error rate for multiclass classification problems, which is not investigated in Guo (2010) and Xu et al. (2015). Simulations and the real data analysis well demonstrate good performances of our methods in comparison with existing methods.     

 

统计学系系列讲座之289

 

时间:2018年4月2日(周一)15:30-16:30

地点:史带楼205室

主持人:朱仲义 教授 复旦大学管理学院统计学系

主题:Fused Mean-variance Filter for Feature Screening

主讲人:唐年胜 教授  云南大学数学与统计学院

简介:唐年胜,博士,国家杰出青年科学基金获得者,教育部“**学者”特聘教授,教育部“新世纪优秀人才”,云南省科技领军人才,云南省首批云岭学者,云南省中青年学术和技术带头人,云南省教学名师,云南省学位委员会经济与管理学科评议组成员,博士生导师。云南省高校“统计与信息技术重点实验室”负责人,“云南大学复杂数据统计推断方法研究”省创新团队带头人。

摘要:A new model-free screening approach called as the slicing fused mean-variance filter is proposed for ultrahigh dimensional data analysis. The new method has the following merits: (i) its implementation does not require specifying a regression form of predictors and response variables; (ii) it can deal with various types of covariates and response variables including continuous, discrete and categorical variables; (iii) it works well even when the covariates/random errors are heavy-tailed, or the predictors are strongly correlated, or there are outliers; (iv) it is unsensitive to the slicing scheme. Under some regularity conditions, the sure screening and ranking consistency properties are established for the proposed procedure without assuming any moment conditions on the predictors. Simulation studies are conducted to investigate the finite sample performance of the proposed procedure. A real data example is illustrated to the proposed procedure.       

 

统计学系系列讲座之290

 

时间:2018年4月2日(周一)16:30-17:30

地点:史带楼205室

主持人:朱仲义 教授 复旦大学管理学院统计学系

主题:A Posterior-Based Wald-Type Statistic for Hypothesis Testing

主讲人:李勇 教授 中国人民大学

简介:李勇,教育部青年**学者,中国人民大学汉青研究院金融学教授,博士生导师, 香港中文大学统计学博士,新加坡管理大学金融学博士后, 中国人民大学量化金融研究所所长。他主要的研究方向是金融计量经济学,量化投资,资产管理。自从2007年12月正式参加工作以来在中英文顶级期刊如《Journal of Econometrics》、《经济研究》、《管理世界》等杂志共发表文章 32篇, 其中SSCI/SCI收录21篇,出版学术专著一部,编著一部。 曾获教育部自然科学二等奖,入选教育部新世纪人才计划,北京市青年优秀人才计划。多家公司战略顾问,证监会,北京市金融局,项目评审专家。

摘要:A new Wald-type statistic is proposed for hypothesis testing based on Bayesian posterior distributions. The new statistic can be explained as a posterior version of Wald test and have several nice properties. First, it is well-defined under improper prior distributions. Second, it avoids Jeffreys-Lindley’s paradox. Third, under the null hypothesis it follows a χ2 distribution asymptotically, offering a pivotal test asymptot­ically. Fourth, it only requires inverting the posterior covariance for the parameters of interest. Fifth and perhaps most importantly, when a random sample from the poste­rior distribution (such as an MCMC output) is available, the proposed statistic can be obtained as a by-product of posterior simulation. In addition, the numerical standard error of the estimated proposed statistic can be computed based on the random sam­ple. The finite sample performace of the statistic is examined in Monte Carlo studies. The method is applied to two latent variable models used in microeconometrics and financial econometrics.     

 

统计学系系列讲座之291

 

时间:2018年3月27日(周二)16:30-17:30

地点:史带楼410室

主持人:朱仲义 教授 复旦大学管理学院统计学系

主题:A Robust t-process Regression Model with Independent Errors

主讲人:Dr Jian Qing Shi School of Mathematics, Statistics & Physics, Cloud Computing for Big Data CDT, University of Newcastle, UK 

摘要:Gaussian process regression (GPR) model is well-known to be susceptible to outliers. Robust process regression models based on t-process or other heavy-tailed processes have been developed to address the problem. However, due to the nature of the current definition for heavy-tailed processes, the unknown process regression function and the random errors are always defined jointly and thus dependently. This definition, mainly owing to the dependence assumption involved, is not justified in many practical problems and thus limits the application of those robust approaches. It also results in a limitation of the theory of robust analysis. In this talk, I will discuss a new robust process regression model enabling independent randomerrors and will also discuss an efficient  estimation procedure. I will present an application to analyse a medical game data and show that the proposed method is robust against outliers and has a better performance in prediction compared with the existing models.

 

统计学系

2018-3-23