Jeffrey Scott McCullough: Empirical Bayesian Estimation
报告时间：2025年7月13日（星期日）19:00-21:30
报告地点：科教楼B座1105
报告人：Jeffrey Scott McCullough 教授
工作单位：University of Michigan
举办单位：经济学院
报告简介：
Bayesian methods have become increasingly vital in modern statistical analysis, offering a coherent framework for incorporating prior knowledge and quantifying uncertainty. This part of the lecture introduces the core principles and practical implementation of Empirical Bayesian Estimation.
It begins with a concise review of Bayesian fundamentals, emphasizing the roles of prior distributions and likelihood functions in deriving posterior estimates. The empirical Bayesian approach is then discussed, where prior distributions are estimated from observed data instead of subjective judgment.
Key topics include selecting appropriate priors, constructing likelihood functions, and integrating data-driven techniques to refine Bayesian inference. Computational strategies such as Markov Chain Monte Carlo (MCMC) and variational inference are highlighted for efficient estimation in complex models. Model selection and validation tools—such as Bayes factors and posterior predictive checks—are also introduced to ensure robustness and reliability.
贝叶斯方法在现代统计分析中变得日益重要，它为整合先验知识与量化不确定性提供了统一而严谨的理论框架。本讲座的第一部分介绍了经验贝叶斯估计的核心原理与实际实现方式。讲座从贝叶斯基本原理出发，强调了先验分布与似然函数在推导后验分布中的关键作用。随后介绍经验贝叶斯方法，其特点是通过观测数据来估计先验分布，而非依赖主观判断。
主要内容包括如何选择合适的先验、构建似然函数，以及如何融合数据驱动技术以提升贝叶斯推断的质量。还将介绍适用于复杂模型的计算策略，如马尔可夫链蒙特卡洛（MCMC）与变分推断，以实现高效估计。此外，本部分也涵盖模型选择与验证方法，包括贝叶斯因子和后验预测检验，以确保推断结果的稳健性和可信度。
报告人简介：
Jeffrey S. McCullough is a Professor of Health Economics at the University of Michigan School of Public Health. His research explores the role of technology and innovation in healthcare, with a particular focus on the economics of information technology, pharmaceuticals, and empirical methods. He has conducted extensive studies on the impact of electronic health record (EHR) systems on healthcare quality and productivity. His current work examines the evolving roles of telemedicine and artificial intelligence in healthcare delivery.
Professor McCullough is a faculty affiliate of the Michigan Institute for Data Science and the Institute for Health Policy and Innovation. He is also a research fellow at the Center for Economic Studies in Munich, Germany. Prior to joining the University of Michigan, he served as an Associate Professor of Health Economics at the University of Minnesota. He earned his undergraduate degree in biochemistry from Harvard College and his Ph.D. in Health Economics from the University of Pennsylvania.
Jeffrey S. McCullough是密歇根大学公共卫生学院的卫生经济学教授。他的研究聚焦于技术与创新在医疗保健中的作用，特别关注信息技术、医药经济学以及实证方法在医疗领域的应用。他曾深入研究电子健康记录系统（EHR）对医疗质量与生产效率的影响。目前，他的研究方向包括远程医疗与人工智能在医疗服务中的不断演变和应用。
Jeffrey S. McCullough教授是密歇根大学数据科学研究院（Michigan Institute for Data Science）和健康政策与创新研究所（Institute for Health Policy and Innovation）的研究成员，同时也是德国慕尼黑经济研究中心（Center for Economic Studies）的研究员。在加入密歇根大学之前，他曾任明尼苏达大学的卫生经济学副教授。他本科毕业于哈佛大学，主修生物化学，并在宾夕法尼亚大学获得卫生经济学博士学位。
