Feature Screening in Ultrahigh-dimensional and Exponentially Tilted likelihood Inference on growing dimensional EEs
报告人: 唐年胜(云南大学)
时间:2024-06-11 10:00-11:00
地点:腾讯会议:660-6689-7741
Abstract:
This talk introduces category-adaptive screening method for heterogeneous categorical data, quantile association-based screening method and slicing fused mean-variance filter for mixed-type data, which are model-free, robust to outliers and heavy-tailed data, and have sure screening properties, ranking consistency property. Also, this talk introduces penalized exponentially tilted likelihood method to estimate parameters in the growing dimensional estimating equations, Wilks theory and the oracle properties are shown under some conditions.
About the Speaker:
云南大学数学与统计学院教授,博士生导师,院长。“国家杰出青年科学基金”获得者,教育部“长江学者特聘教授”,国家级教学名师,教育部“新世纪优秀人才支持计划”入选者,国家百千万人才工程暨有突出贡献中青年专家,享受国务院政府特殊津贴。国际统计学会推选会员,国际数理统计学会会士(IMS Fellow)。现兼任教育部高等学校统计学类教学指导委员会委员、中国统计学会常务理事、中国现场统计研究会副理事长等;Statistics and Its Interface、Communications in Mathematics and Statistics 等杂志的编委或副主编。研究方向包括:生物统计、贝叶斯统计、缺失数据分析、高维数据分析、大数据统计等,在JASA、Annals of Statistics、Biometrika、Journal of Econometrics、JMLR等学术期刊发表论文200余篇;出版学术专著5部、译著2部、教材1部。
Your participation is warmly welcomed!
欢迎扫码关注北大统计科学中心公众号,了解更多讲座信息!