欢迎来到永利集团304am登录!
首页
学术动态
首页 > 正文

【预告】北京工业大学李高荣教授学术报告

来源: 日期:2017-09-20 作者: 浏览次数:

报告题目: 

1、Profile forward regression screening forultra-high dimensional semiparametric varyingcoefficient partially linear models 

2、SIMEX estimationfor single-index model with covariate measurement error 

报告时间: 

1、2017年9月21日下午16:00 

2、2017年9月22日上午9:00 

报告地点: 

1、科学会堂A602 

2、科学会堂A710 

报告摘要: 

1、In this paper, we consider semiparametric varyingcoefficient partially linear models when the predictor variables of the linearpart are ultra-high dimensional where the dimensionality grows exponentiallywith the sample size. We propose a profile forward regression (PFR) method toperform variable screening for ultra-high dimensional linear predictorvariables. The proposed PFR algorithm can not only identify all relevantpredictors consistently even for ultra-high semiparametric models includingboth nonparametric and parametric parts, but also possesses the screeningconsistency property. To determine whether or not to include the candidatepredictor in the model of selected ones, we adopt an extended Bayesianinformation criterion (EBIC) to select the ‘‘best’’ candidate model. Simulationstudies and a real data example are also carried out to assess the performanceof the proposed method and to compare it with existing screening methods.  

2、In this paper, weconsider the single-index measurement error model with mismeasured covariatesin the nonparametric part. To solve the problem, we develop asimulation-extrapolation (SIMEX) algorithm based on the local linear smootherand the estimating equation. For the proposed SIMEX estimation, it is notneeded to assume the distribution of the unobserved covariate. We transform theboundary of a unit ball to the interior of a unit ball by using the constraint$\|\beta\|=1$. The proposed SIMEX estimator of the index parameter is shown tobe asymptotically normal under some regularity conditions. We also derive theasymptotic bias and variance of the estimator of the unknown link function. Finally,the performance of the proposed method is examined by simulation studies and isillustrated by a real data example. 

报告人简介: 

李高荣,北京工业大学教授,博士生导师。主要研究方向是复杂高维数据分析、深度学习、模型和变量选择、非参数统计、经验似然、纵向数据和测量误差模型等。于2007年7月在北京工业大学应用数理学院获得概率论与数理统计专业博士学位,2007年8月到2009年6月在华东师范大学金融与统计学院从事博士后研究工作,2016年3月到2017年3月为美国南加州大学Marshall商学院博士后。迄今为止,在《The Annals of Statistics》、《Statisticsand Computing》、《Journal of Multivariate Analysis》、《Statistica Sinica》和《ComputationalStatistics and Data Analysis》等国内外重要学术期刊发表学术论文70多篇,其中40多篇发表在国际SCI期刊,在科学出版社出版专著《纵向数据半参数模型》和《现代测量误差模型》。目前主持国家自然科学基金、北京市自然科学基金和北京市教委科技计划面上项目。入选北京市属高等学校人才强教深化计划“中青年骨干人才培养计划”和北京市优秀人才培养资助计划和北京工业大学“京华人才”。