A New Bayesian Strength of Evidence for Testing a Point Null Hypothesis

内容来源: 中国人民大学统计学院

题目:A New Bayesian Strength of Evidence for Testing a Point Null Hypothesis
报告日期:2020-11-02,9:00
报告地点:腾讯会议(会议ID:509 708 658)

报告人简介
汪敏(Min Wang),美国德州大学圣安东尼奥分校 (University of Texas at San Antonio) 商学院管理科学与统计系副教授(获终身教职),博士生导师。2010年5月于美国克莱姆森大学(Clemson University)获得统计硕士学位;2013年5月于克莱姆森大学大学获得统计博士学位。2013年8月- 2017年12月在美国密歇根理工大学数学科学系工作和在2017年8月破格提前提升为副教授并获得终身任期教授资格;现在在德州大学圣安东尼奥分校从事教学科研工作。近年来,先后参与和主持了美国自然科学基金委(NSF),密歇根交通部,以及美国卫生院(NIH)的研究课题。在各类同行评议的国际权威期刊上发表了研究文章50余篇。研究方向:贝叶斯统计;计算统计;统计推断;质量和可靠性工程研究;高维数据分析和统计应用。

报告摘要
The frequentist evidence expressed in terms of the observed level of significance, and the Bayesian evidence expressed through the posterior probability and the Bayes factor, are two main statistical streams of thought for testing a point null hypothesis. However, they may give rise to different decisions in practical situations and could even cast serious doubt on the adequacy of the two schools of evidence for hypothesis testing. In this talk, we propose a new Bayesian strength of evidence, which can not only reconcile the disagreement between frequentists and Bayesians in many classical examples in which Lindley’s paradox occurs,but also is shown to be a Bayes test under some specific loss functions. Thus, it can be viewed as an objective and automatic Bayesian approach to hypothesis testing. Finally, two applications are provided for illustrative purposes.

发表评论

邮箱地址不会被公开。 必填项已用*标注