Advancing spatial analysis methodologies using a bayesian approach: combining individual and aggregated data in small area studies

使用贝叶斯方法推进空间分析方法：在小区域研究中结合个体数据和聚合数据

• 批准号: 371625-2009
• 负责人: Law, Jane
• 金额: $ 1.68万
• 依托单位: University of Waterloo
• 依托单位国家: 加拿大
• 项目类别: Discovery Grants Program - Individual
• 财政年份: 2012
• 资助国家: 加拿大
• 起止时间: 2012-01-01 至 2013-12-31
• 项目状态: 已结题
• 来源: https://www.nserc-crsng.gc.ca/ase-oro/Details-Detailles_eng.asp?id=505093
• 关键词: Advancing; spatial; analysis; methodologies; using

The proposed research is important to all disciplines that use spatial data because it will develop new and improved methodologies for analyzing spatial data, and thus make effective use of spatial data that have been collected routinely across disciplines. Research that applies spatial analysis has increased remarkably in many disciplines as more geographically-referenced data are becoming available. Traditional spatial analysis adopts a frequentist statistical approach, which has the limitation of inability to fit complex spatial models. One important development of spatial analysis in recent years is to use a Bayesian statistical approach to analyze spatial data, referred to a Bayesian Spatial Analysis (BSA). Research has shed lights on the capability of this new approach in fitting more complex spatial models to tackle traditional problems of spatial analysis including missing data, data uncertainty, measurement errors, spatial dependence, and data integration. Compared to the UK and US, BSA is rarely used in Canada. The long term goal of our proposed research is to advance methodologies of BSA to improve our ability to analyze geographic data reliably, provide BSA training to students, promote the applications of BSA, and improve the overall standard of geographic research in Canada. Our short term goal is to develop BSA methodologies that would enable the integration of geographically-referenced individual and aggregated data in small-area research to improve the reliability of geographic studies. We will focus on developing methods for research in the fields of public health and crime, which are currently most in need of improved spatial methodologies for understanding geographic variation of disease/crime outcomes. Our research findings will be shared with two regions of public health and police in Ontario who have expressed interests in and agreed to provide data for our research. However, the BSA methodologies developed will be applicable not only in public health and crime, but also in environmental science and other applied science research.

拟议的研究对于使用空间数据的所有学科都很重要，因为它将开发出新的和改进的方法来分析空间数据，从而有效利用跨学科常规收集的空间数据。 随着越来越多的地理参考数据的可用性，应用空间分析的研究在许多学科中都显着增加。 传统的空间分析采用了一种常见的统计方法，该方法限制了不适合复杂空间模型的局限性。 近年来，空间分析的一个重要发展是使用贝叶斯统计方法分析空间数据，该方法称为贝叶斯空间分析（BSA）。 研究已经阐明了这种新方法拟合更复杂的空间模型的能力，以解决传统的空间分析问题，包括缺少数据，数据不确定性，测量误差，空间依赖性和数据集成。 与英国和美国相比，BSA在加拿大很少使用。 我们拟议的研究的长期目标是提高BSA的方法，以提高我们可靠地分析地理数据，向学生提供BSA培训，促进BSA的应用并提高加拿大地理研究的整体标准的能力。 我们的短期目标是开发BSA方法，以使小区域研究中的地理参考个体和汇总数据整合，以提高地理研究的可靠性。 我们将专注于开发公共卫生和犯罪领域研究的方法，目前最需要改进的空间方法来理解疾病/犯罪结果的地理变化。 我们的研究发现将与安大略省的两个公共卫生和警察区域共享，他们表达了利益并同意为我们的研究提供数据。 但是，开发的BSA方法不仅适用于公共卫生和犯罪，还适用于环境科学和其他应用科学研究。