云亭数学讲坛2022第80讲—— 田茂再教授

文章来源:bat365在线平台官网登录发布日期:2022-11-09浏览次数:220

应学院邀请,中国人民大学田茂再教授将在线作学术报告。

报告题目:Censored Quantile Transformation Models and Its

Applications

报告摘要:Alzheimer's disease (AD) is a progressive disease that starts from mild cognitive impairment and may finally lead to memory loss. Therefore, it is critical to explore the risk factors for the conversion time to AD. However, previous studies in this context mainly focus on modeling single functional covariate, leaving out many important information. In our motivating AD data, both the left and right hippocampal radial distance curves are potentially significant, but they are correlated with each other, which may cause serious multicollinearity in the regression. Thus, we propose a multivariate functional censored quantile regression model based on the Box-Cox transformation to analyze the right-censored data. The multiple functional predictors are jointly handled by the multivariate functional principal component analysis. By introducing dynamic power transformations, the proposed method relaxes the global linear assumption imposed by many existing studies and enjoys more flexibility in the framework of censored quantile regression. Based on the martingale theory, uniform consistency and weak convergence are established as a process of quantile levels. Simulation studies demonstrate the outperformance of the proposed method. Real data analysis shows the importance of both left and right hippocampal radial distance curves for predicting the conversion time to AD in different quantile levels. Supplementary materials for this talk are available online.

报告时间:2022111308:00

报告地点:腾讯会议号 (371267370)

邀 请 人:田玉柱副教授 肖鸿民教授

届时欢迎广大师生参与交流!


报告人简介

 田茂再,二级教授,博士研究生导师。他先后在中国科学院、加拿大艾伯塔大学、卡尔加里大学、香港中文大学、香港浸会大学、澳大利亚墨尔本大学做过6个博士后,是德国洪堡大学SFB 649 FELLOW重大科研项目中方首席科学家,是美国耶鲁大学、哥伦比亚大学、英国曼彻斯特大学、布鲁奈尔大学、日本东京大学以及意大利佛罗伦萨大学的高级访问教授,曾经入选新疆医科大学高层次人才、新世纪优秀人才、甘肃省“飞天学者”和兰州财经大学“兴隆学者”特聘教授、新疆维吾尔自治区“天山学者”特聘教授,中国人民大学首批杰出学者,新疆维吾尔自治区“天池学者”特聘教授,以及中国人民大学应用统计科学研究中心研究员。他先后主持省部级、国家级项目40项左右,在国内外的重要学术刊物上发表近300篇颇具影响的文章,著书15部,获省部级及以上奖励10余项。他的科研业绩重点是复杂大数据的统计建模理论,这些研究成果产生了大量的后续他人跟踪研究,极大地推动了统计学科的进步,并影响了其他相关学科的发展,在实践领域发挥着重要作用,其创新程度高,在国际上得到了同行专家的广泛认可。其学术兼职30余个,包括中国概率统计学会理事,国际生物统计学会中国分会 (IBS-China) 常务理事,中国现场统计研究会高维数据统计分会常务理事,国家自然科学基金同行评议专家,国家社会科学基金同行评议专家,英国皇家统计学会(JRSS)等国际国内近百家学术期刊审稿人,美国数学会的著名杂志《数学评论》(Mathematical Reviews)评论员。