Seventh International Workshop on Objective Bayesian Methodology; Philadelphia, PA

第七届客观贝叶斯方法论国际研讨会；

• 批准号: 0924257
• 负责人: Lawrence Brown
• 金额: --
• 依托单位: University of Pennsylvania
• 依托单位国家: 美国
• 项目类别: Standard Grant
• 财政年份: 2009
• 资助国家: 美国
• 起止时间: 2009-07-01 至 2010-06-30
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=0924257&HistoricalAwards=false
• 关键词: Seventh; International; Workshop; Objective; Bayesian

The investigator will coordinate a five-day Objective Bayes workshop to be held at the Wharton School of the University of Pennsylvania. This workshop will bring together leading researchers from around the world who are active in the area of objective prior methodology. It may be viewed as a continuation of previous international meetings convened over the past 12 years devoted to Objective Bayes theory and methodology. The main objectives of the meeting are to facilitate the exchange of recent research developments within this scientific community, to provide opportunities for new researchers and underrepresented groups to be knowledgeable and active in this important area of research, and to establish new collaborations that will channel efforts into pending problems and open new directions for investigation. Objective Bayesian methodology is, for the most part, oriented towards the development of prior distributions that can be used automatically, i.e., that do not require subjective input other than the specific probabilistic model chosen to describe the data. Accompanying the development of forms particularly suitable for particular types of applications is the study of computational techniques for their implementation and evaluation of both the theoretical and concrete implications of their use. Many of the topics proposed for the workshop reflect an emphasis on objective Bayesian methodology for general classes of practical applications such as spatial-temporal models, hierarchical random-effects models, multiple comparisons and goodness -of-fit. In the direction of specific applications there will be sessions on Bayesian applications in astrophysics and on Bayesian applications in Business and Marketing research. In a more theoretical direction, there will be sessions on the foundations of objective Bayes analysis, including a special retrospective examination of Harold Jeffreys' seminal contributions to the area, and on the unification of Bayesian and frequentist statistical methods. The principal behind these so-called objective Bayes methods also has a long history dating at least back to work of Laplace almost 200 years ago. But two much more recent developments have brought emphasis and importance to the use of these methods. One is the rapidly increasing massiveness and complexity of modern statistical data. This has put extreme and often unmanageable strains on the theory and implementation of conventional non-Bayesian methods. (At the same time such complexity almost completely denies any reasonable possibility of formulating reasonable subjective prior opinions needed for ordinary Bayesian analyses.) The other parallel factor is the development within the past two decades of a suite of mathematical and computational strategies for successfully computing appropriately structured objective Bayesian procedures in highly complex situations. Within this time span an expanding community of scholars has been investigating these objective Bayes models and methods. The current conference proposal is to organize a workshop devoted to the investigation of objective Bayes methods.

研究人员将协调在宾夕法尼亚大学沃顿商学院举行的为期五天的客观贝叶斯研讨会。该研讨会将汇集来自世界各地的领先研究人员，他们活跃在客观先验方法学领域。它可以被看作是过去12年来专门讨论客观贝叶斯理论和方法的国际会议的延续。会议的主要目标是促进在这一科学界内交流最新的研究进展，为新的研究人员和代表性不足的群体提供机会，使他们在这一重要的研究领域具有知识和积极性，并建立新的合作关系，将努力引导到悬而未决的问题上，并为调查开辟新的方向。客观贝叶斯方法在很大程度上是面向开发可以自动使用的先验分布，即，除了选择用于描述数据的特定概率模型之外，不需要主观输入。伴随着特别适合于特定类型的应用程序的形式的发展，是他们的实施和评估其使用的理论和具体影响的计算技术的研究。为讲习班提出的许多专题反映了对一般类别的实际应用，如时空模型、分层随机效应模型、多重比较和拟合优度的客观贝叶斯方法的重视。在具体应用的方向将有会议上贝叶斯应用天体物理学和贝叶斯应用在商业和营销研究。在一个更理论的方向，将有会议的基础上客观贝叶斯分析，包括一个特殊的回顾性审查哈罗德杰弗里斯的开创性贡献的领域，并统一贝叶斯和频率统计方法。这些所谓的客观贝叶斯方法背后的原则也有很长的历史，至少可以追溯到近200年前的拉普拉斯的工作。但最近的两个发展使这些方法的使用得到强调和重视。一个是现代统计数据的复杂性和复杂性迅速增加。这给传统的非贝叶斯方法的理论和实现带来了极端的、往往是难以管理的压力。(At与此同时，这种复杂性几乎完全否定了任何合理的可能性，制定合理的主观先验意见所需的普通贝叶斯分析。另一个平行因素是在过去二十年的一套数学和计算策略的发展，成功地计算适当结构的客观贝叶斯程序在高度复杂的情况下。在这段时间内，越来越多的学者一直在研究这些客观的贝叶斯模型和方法。目前的会议建议是组织一个专门研究客观贝叶斯方法的研讨会。