5月28日上海海事大学范国良教授来我院讲座预告

发布者:王玲芳发布时间:2021-05-24浏览次数:4


讲座主题: Weighted empiricallikelihood for heteroscedastic varying coefficient partially nonlinear modelswith missing data

主讲人:范国良

讲座时间:2021年5月28日下午15:40-16:30

地点:综合楼644

主讲人简介:    



范国良,上海海事大学经济管理学院教授,2010年博士毕业于同济大学,中国人民大学博士后,中国现场统计研究会资源与环境统计分会理事,中国工程概率统计学会理事,中国商业统计学会理事。主要研究领域包括:充分降维理论、分位数回归方法、经验似然等。在《Statistica Sinica》、《Electronic Journal ofStatistics》、《Journal of Multivariate Analysis》、《Journal of Statistical Planning and Inference》、《Science China Mathematics》等国内外学术刊物上发表学术论文四十余篇;主持国家自然科学基金、教育部人文社会科学研究项目、上海市自然科学基金、中国博士后科学基金、中国博士后科学基金特别资助、安徽省自然科学基金、安徽省教育厅重点项目等项目十多项,曾获安徽省科学技术奖以及安徽省自然科学优秀学术论文奖。                   

讲座摘要:


In this talk, aweighted empirical likelihood technique for constructing the empiricallikelihood confidence regions is applied to study the heteroscedastic varyingcoefficient partially nonlinear models with missing response data. We firstgive the estimator of the error variance based on the Nadaraya-Watson kernelestimation method. Then a weighted empirical log-likelihood ratio of theunknown parameter is constructed based on the inverse probability weightedtechnique. The maximum empirical likelihood (MEL) estimator of the unknownparameter is obtained. Further, a weighted empirical log-likelihood ratio ofthe varying coefficient function is introduced based on the MEL estimator andthe inverse probability weighted method. The limiting distributions of theresulting statistics both for the unknown parameter and varying coefficientfunction are shown to have standard chi-squared distribution. A simulationstudy and a real data set are undertaken to investigate the finite sampleperformance of the proposed methods.