Mathematical Sciences: Bayesian Experimental Design and Prior Elicitation in Response to Surface Models

数学科学：贝叶斯实验设计和对表面模型的先验启发

• 批准号: 9209378
• 负责人: Blaza Toman
• 金额: $ 1.78万
• 依托单位: George Washington University
• 依托单位国家: 美国
• 项目类别: Standard Grant
• 财政年份: 1992
• 资助国家: 美国
• 起止时间: 1992-07-15 至 1994-12-31
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=9209378&HistoricalAwards=false
• 关键词: Mathematical; Sciences; Bayesian; Experimental; Design

This project begins the preliminary research in the development and implementation of optimal Bayesian experimental designs based on expert opinion and informative priors. Expansions from fixed designs derived from fractional factorial designs will include sequential designs for response surfaces and designs for yield optimization. The role and practicalities of prior elicitation processes will also be examined. Bayesian statistical methodology allows direct incorporation of expert opinion or of good prior information, as is often available for experiments in physical science and manufacturing. This project will develop Bayesian optimal designs for experiments for response surface models, for yield optimization and for particular application with the Taguchi method.

本项目在国内率先开展前期研究， 最优贝叶斯实验的开发与实现 基于专家意见和信息先验的设计。 扩张 从部分析因设计衍生的固定设计将 包括响应面顺序设计和 产量优化 事先知情同意的作用和实用性 还将审查启发过程。 贝叶斯统计方法允许直接合并 专家意见或良好的先验信息，因为通常 可用于物理科学和制造业的实验。 本计画将发展贝氏最佳实验设计 用于响应面模型、产量优化和 特别是应用田口方法。