讲座题目：Sequential adaptive variables and subject selection for GEE methods

主讲人：王占锋  中国科学技术大学

讲座时间：2019年12月16日（周一）下午3：30—4：30

讲座地点：综合楼601

主讲人简介：

王占锋，中国科学技术大学统计与金融系副教授，概率与统计教研室主任和应用统计专业硕士学位项目主管。分别于2003年和2008年获中国科学技术大学学士和理学博士学位。主要从事

统计渐近理论，生物统计，函数型数据分析等领域的研究，在国内外学术期刊上发表和接受论文40篇。曾主持国家自然科学青年基金一项和自然科学基金一项，参与国家重点自然科学基金一项和自然科学基金一项。中国现场统计研究会资源与环境统计分会理事，全国工业统计学研究会理事和中国青年统计学家协会常务理事。

讲座摘要：

Modeling correlated or highly stratified multiple-response data is a common data analysis task in many applications such as in large epidemiological studies or multi-site cohort studies. Generalized estimating equations (GEE) is one of the popular statistical methods for analyzing this kind of data, because they can handle many types of unmeasured dependence between outcomes. Collecting large number of the highly stratified or correlated responses data is a time-consuming job. Thus, to have a more aggressive sampling strategy that can accelerate this process, as active learning methods in machine learning literature,  is always beneficial.   In this paper, we integrate the adaptive sampling and variable selection features into a sequential procedure for modeling the correlated responses data.  Besides reporting the statistical properties of the proposed procedure, we also use both synthesized data and real data sets to demonstrate the usefulness of our method.