报告题目：On penalized jackknife empirical likelihood with a diverging number of U estimating equations
时 间：2022年3月23日（星期三）下午14：30-16：00
地 点：腾讯会议（928-630-468）
主 办：数学与统计学院
参加对象：感兴趣的老师和学生
报告摘要：Jackknife empirical likelihood is an attractive approach for statistical inferences with nonlinear statistics such as U-statistics. However, most contemporary problems involve high dimensional model selection and its feasibility in theory and practice remains largely unexplored in the situations where the number of parameters diverges to infinity. In this paper, we propose a penalized JEL method which preserves the main advantages of JEL and leads to reliable variable selection based on the estimating equations with U-statistic structure in the high-dimensional setting. Under certain regularity conditions, we establish the asymptotic theory and oracle property for the JEL and its penalized version when the number of estimating equations and parameters increases along with the sample size. Simulation studies and real data analysis were carried out to examine the performance of the proposed methods and illustrate its practical utilities.
报告人简介：徐锦峰，本科和硕士毕业于中国科学技术大学，于美国哥伦比亚大学获统计学博士学位，现任职于香港大学统计及精算学系。主要研究领域为生存分析，高维数据，生物统计和机器学习。在《Journal of the American Statistical Association》、《Biometrika》、《Journal of Econometrics》、《Biometrics》、《Scandinavian Journal of Statistics》、《Statistics in Medicine》、《AISM》统计学、经济学与医学等领域重要期刊上发表过七十余篇的学术论文，先后主持香港政府大学研究基金资助局项目三项。