统计学系系列讲座之354期

时间:2019年9月26日(星期四)17:00-17:30

地点:史带楼205室

主持人:张新生  教授 复旦大学管理学院统计学系

主题:It’s the interaction, stupid

主讲人:Prof. Heping Zhang

Susan Dwight Bliss Professor of Biostatistics, Yale University School of Public Health

简介:

张和平博士,耶鲁大学Susan Dwight Bliss生物统计学教授,统计与数据科学教授,儿童研究中心教授,创建并主持耶鲁大学科学与统计协作中心, 担任期刊Statistics and Its Interface的创始主编,目前担任美国统计协会杂志(JASA), 遗传流行病学和生殖与不育专题研究的编委。2019担任JASA (ACS)主编。张教授曾入选哈佛大学公共卫生学院2008年度Myrto Lefkopoulou杰出学者并作2011年IMS Medallion报告,2011年Royan国际生殖健康研究奖的获得者,2013年获得美国生殖医学学会颁发的科学论文奖,2014年March of Dimes 早产最佳研究奖,2017年美国妇产科杂志优秀论文奖。张教授的研究兴趣包括非参数方法,纵向数据,统计遗传学和生物信息学,临床试验,流行病学数据统计建模,脑成像分析,统计计算和行为科学的统计方法。出版著作“递归分区及其应用(Recursive Partitioning and Its Applications)”,并在高影响力的统计、遗传、流行病学和精神病学期刊上发表了290多篇学术论文,其中包括Annals of Statistics, Annals of Applied Statistics, Biometrika, JASA, JRSSB, American Journal of Human Genetics, American Journal of Psychiatry, PNAS, Science, JAMA, 以及 the New England Journal of Medicine。

摘要:

The vast majority of statistical methods, theory, and applications are for or based on additive models, and often linear models. Those models have led to reasonable and easily interpretable models. However, in practice, more likely than not, the underlying data structures are not additive, and the only difference is to what degree the additive models provide a sufficiently good fit to the data. For example, in the analysis of data from genomewide association studies, the failure to identify genes with major effects on complex diseases is very likely due to our inability to consider and identify gene-gene and/or gene-environment interactions. In this talk, I will present two approaches to detecting interactions and demonstrate their potential with numerical examples.