Prior-free Probabilistic Inference: Inferential Models

报告摘要：As the tool for the science that converts experience, in the form of observed data, to knowledge about unknown quantities of interest, Statistics will be fundamental to the ultimate success of “data science”. Developing solid foundations for scientific inference is the most fundamental but unsolved problem in statistics. We argue for two new basic principles, namely, the validity and efficiency principles, for truly prior-free probabilistic inference. With a brief introduction to a principle-based framework, called Inferential Models (IMs), this talk focuses on demonstrating how IMs can provide deep understanding of prior-probabilistic inference for combining information and parameters of interest, called Conditional IMs and Marginal IMs.

个人简介： 普渡大学统计系正教授。1994年毕业于哈佛大学统计系，博士学位。刘传海教授是国际计算和统计推断学方面的专家，在国际统计学刊物发表逾70余篇，发表英文专著两本。刘传海教授是：美国统计学会（ASA）FELLOW和国际统计协会（ISI）会员。现/历任： Journal of the American Statistical Association， Statistica Sinica，Journal of Computational and Graphical Statistics， Journal of Multivariate Analysis等国际知名统计学杂志副主编（Associate Editor）。在业界实践方面，刘传海教授有着十余年在贝尔实验室工作的经验，是国际上最早关注并研究大数据的大学学者, 在此期间积累了丰富的统计软件平台开发经验。