Bayesian Analysis, Computation and Communication in the Social Sciences

社会科学中的贝叶斯分析、计算和传播

• 批准号: 9731037
• 负责人: John Geweke
• 金额: $ 23.99万
• 依托单位: University of Minnesota-Twin Cities
• 依托单位国家: 美国
• 项目类别: Continuing Grant
• 财政年份: 1998
• 资助国家: 美国
• 起止时间: 1998-07-01 至 2001-06-30
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=9731037&HistoricalAwards=false
• 关键词: Bayesian; Analysis; Computation; Communication; Social

The objective of this project is to provide tools for Bayesian inference that will be used by social scientists and their students - not just specialists in statistics or methodology, but anyone who does quantitative work with data, including undergraduates. This is an extension of a previous project that developed tools for Bayesian communication. These tools allow the user to approximate posterior moments, modify prior distributions, combine models, and approximate posterior density functions, given the output of a posterior simulator, which can be made available by the investigator at an ftp or web site. The proposed project extends this work by placing posterior simulators themselves at the disposal of social scientists, in an environment that meets current standards for user orientation. Ongoing research will support the following steps to meet this objective: 1) Develop redundant algorithms for the models that form the analytic core of empirical work in the social sciences. In large part, a generic Hastings-Metropolis algorithm will provide this redundancy. 2) Provide windowing interfaces for all procedures including posterior simulators, the Bayesian communication tools developed in the previous project, and new procedures to be developed in this project. Make these interfaces available for all common operating systems, using the scripting language and windowing toolkit Tcl/Tk. 3) Facilitate incorporation of new models and methods in the software. The proposed system is designed to incorporate simulators developed by third parties, with no modification of their code. All the software developed in this project will be freely available and fully documented for anyone with an internet connection.

这个项目的目标是提供贝叶斯推理的工具，这些工具将被社会科学家和他们的学生使用-不仅仅是统计学或方法学专家，而是任何对数据进行定量工作的人，包括本科生。 这是以前开发贝叶斯通信工具的项目的扩展。 这些工具允许用户近似后验矩，修改先验分布，联合收割机模型，并近似后验密度函数，给定后验模拟器的输出，研究者可以在ftp或网站上提供。 拟议中的项目扩展了这项工作，将后验模拟器本身放置在社会科学家的处置，在一个环境中，满足当前的用户导向标准。 正在进行的研究将支持以下步骤来实现这一目标：1）为形成社会科学实证工作分析核心的模型开发冗余算法。 在很大程度上，通用的Hastings-Metropolis算法将提供这种冗余。 2)为所有程序提供窗口界面，包括后验模拟器、先前项目中开发的贝叶斯通信工具以及本项目中待开发的新程序。 使用脚本语言和窗口工具包Tcl/Tk，使这些接口可用于所有常见的操作系统。 3)促进软件中新模型和方法的整合。 拟议的系统旨在纳入第三方开发的模拟器，而不修改其代码。 在这个项目中开发的所有软件都将免费提供，并为任何有互联网连接的人提供完整的文档。