Bayesian Inference and Econometric Modeling

贝叶斯推理和计量经济建模

• 批准号: 9514382
• 负责人: Arnold Zellner
• 金额: $ 19万
• 依托单位: University of Chicago
• 依托单位国家: 美国
• 项目类别: Continuing Grant
• 财政年份: 1995
• 资助国家: 美国
• 起止时间: 1995-10-01 至 2000-09-30
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=9514382&HistoricalAwards=false
• 关键词: Bayesian; Inference; Econometric; Modeling

The Bayesian Method of Moments (BMOM) approach to statistical and econometric estimation, prediction, and model formulation allows to the researcher to obtain parameter estimates without the distributional assumptions required for traditional Bayesian analysis. The objectives of this research project are (1) to develop further and provide additional applications of the BMOM approach; (2) to compare BMOM and traditional Bayesian and non-Bayesian inference techniques and develop procedures for choosing between and/or combing BMOM and traditional Bayesian results; and (3) to utilize BMOMs and other procedures in applied model formulation and forecasting applications involving multivariate dynamic statistical and econometric models useful in gaining understanding of past experience and in point and turning point forecasting. Applications will involve modeling and forecasting output growth rates for 18 or more industrialized economies and analyses of U.S. regional petroleum markets. Past experience has indicated and it is anticipated that new experience with the BMOM approach will indicate that it is operational, simpler than previous approaches, and yields very satisfactory results.

贝叶斯矩量法（BMOM）用于统计和计量经济学估计、预测和模型制定的方法使研究人员无需传统贝叶斯分析所需的分布假设即可获得参数估计。本研究计划的目标是：(1)进一步发展和提供BMOM方法的其他应用；(2)比较BMOM与传统贝叶斯和非贝叶斯推理技术，并制定在BMOM与传统贝叶斯结果之间进行选择和/或结合的程序；(3)在涉及多元动态统计和计量经济模型的应用模型制定和预测应用中利用bmom和其他程序，这些模型有助于了解过去的经验，并有助于进行点和转折点预测。应用将包括对18个或更多工业化经济体的产量增长率进行建模和预测，以及对美国地区石油市场的分析。过去的经验表明，预期BMOM方法的新经验将表明，它是可操作的，比以前的方法更简单，并产生非常令人满意的结果。