时 间：2021年11月18日（周四） 10：00-11：30
地 点：线上，腾讯会议：685 263 364
题 目：Density ratio model with data-adaptive basis function
主讲人：陈家骅 云南大学&英属哥伦比亚大学教授
主持人：刘玉坤 教授
摘 要：
In many applications, we collect samples from interconnected populations. These population distributions share some latent structure, so it is advantageous to jointly analyze the samples to enhance statistical efficiency. One effective way to connect the distributions is the density ratio model (DRM). A key ingredient in the DRM is that the log density ratios are linear combinations of pre-specified functions; the vector formed by these functions is called the basis function. The benefit of DRM, however, relies on correctly specifying the basis function. In applications, we do not have complete knowledge to enable a perfect choice of the basis function. A data-adaptive choice of the basis function can alleviate the risk of model misspecification, and it remains an open problem. In this talk, we discuss a data-adaptive approach to the choice of basis function based on functional principal component analysis (FPCA). Under some conditions, we show that this approach leads to consistent basis function estimation. Our simulation results show that the proposed adaptive choice leads to an efficiency gain. We use a house income data set to demonstrate the efficiency gain and the ease of our approach.
报告人简介：
陈家骅，加拿大英属哥伦比亚大学（UBC）统计系国家一级讲座教授，云南大学大数据研究院院长。曾任泛华统计学会主席、加拿大统计杂志主编等职务。1983年本科毕业于中国科大数学系，1985年硕士毕业于中国科学院系统科学研究所，1990年于美国威斯康星大学麦迪逊分校统计学系获得博士学位，师从吴建福教授。研究兴趣包括混合模型、试验设计、经验似然、大样本理论和变量选择等多个统计研究领域，在顶级统计学期刊如JASA, JRSSB, Annals of Statistics, Biometrika等上发表论文100多篇。曾获多项学术荣誉：2005年被加拿大统计学会授予CRM-SSC年度奖；2005年当选fellow of the Institute of Mathematical Statistics；2009年当选fellow of the America Statistical Associate；2014年获加拿大统计学会最高金奖；2016年获泛华统计协会杰出成就奖。