Mathematical Sciences: Workshop on Bayesian Statistics in Science and Technology

提议：DMS 9505461 PI：罗伯特·卡斯，康斯坦丁·盖特森西，詹姆斯·霍奇斯，诺泽尔·辛普尔瓦拉研究所：卡内基梅隆大学标题：贝叶斯案例研究研讨会，1995年10月摘要：在宾夕法尼亚州匹兹堡卡内基-梅隆大学举行的题为‘科学和技术中的贝叶斯统计：案例研究三’的专题讨论会，包括贝叶斯方法应用的扩展介绍。这些应用所涉及的问题中，统计学家是研究团队中不可或缺的一员。专题讨论会的目的是：(1)确定并集中注意妨碍应用贝叶斯方法的具体实施和理论问题，并确定候选解决方案；(2)提供一个论坛，在具体研究项目的背景下探讨统计理论与实践之间的相互作用；(3)提供一个小型会议氛围，使初级研究人员和研究生能够与有经验的研究人员一起探索实质性的贝叶斯应用；以及(4)编写一册文件齐全的案例研究和数据集，供应用统计和其他量化领域的研究人员、从业人员、教育工作者和学生使用。今天，进行调查的科学家可以获得越来越多的背景信息。在设计研究和分析数据时，有效地利用先前的知识是很重要的。贝叶斯统计方法就是为此量身定做的。贝叶斯统计理论和计算最近取得了许多进展，但科学会议很少花大量时间讨论这些方法在实质性科学问题上的成功应用。本次研讨会只关注贝叶斯统计的应用。应用领域从经济学到种群生物学。其结果将是阐明理论和实践之间的相互作用，并确定成功的方法，这些方法将为未来的研究指明重要方向。讲习班的互动和动态将通过录像带和更传统的会议记录书面形式记录和传播。

Proposal: DMS 9505461 PIs: Robert Kass, Constantine Gatsonsi, James Hodges, Nozer Singpurwalla Institution: Carnegie Mellon University Title: Bayesian Case Studies Workshop, October 1995 Abstract: The symposium entitled ``Bayesian Statistics in Science and Technology: Case Studies III'' held at Carnegie-Mellon University in Pittsburgh, Pennsylvania, includes extended presentations of applications of Bayesian methods. These applications are in problems in which the statistician is an integral member of the research team. The objectives of the symposium are to: (1)identify and focus attention on specific implementation and theoretical problems that hinder applications of Bayesian methods, and to identify candidate solutions; (ii) provide a forum in which the interplay between statistical theory and practice will be explored in the context of concrete research projects; (iii) provide a small-meeting atmosphere within which junior investigators and graduate students can explore substantial Bayesian applications with experienced researchers; and (iv) produce a volume containing well-documented case studies and data sets suitable for use by researchers, practitioners, educators and students of applied statistics and other quantitative fields. Today there are increasing amounts of background information available to scientists undertaking an investigation. It is important to utilize previousknowledge effectively in designing studies and analyzing data. Bayesian statistical methods are tailored to this purpose. There have been many recent advances in Bayesian statistical theory and computation, but scientific meetings rarely spend substantial time discussing successful applications of these methods to substantive scientific problems. This symposium concentrates attention solely on applications of Bayesian statistics. The application areas range from economics to population biology . The result will be the elucidation of the interplay between theory and practice and the identification of successful methods that will point to important directions for future research. The interactions and dynamics of the workshop will be captured and disseminated via video tape and in the more traditional written form of a conference proceedings.