Geostatistical Software for Non-Parametric Geostatistical Modeling of Uncertainty

7.项目摘要/摘要 在对两国之间的关联和/或因果关系进行任何调查中的关键组成部分 环境和健康结果是提供准确和精确的暴露模型。因为 收集外业数据的成本通常令人望而却步，因此整合任何二次信息来源至关重要 可用于补充稀疏数据集。辅助数据可以采用多种形式(例如，连续的或分类的 测量尺度)，并显示不同级别的可靠性：硬数据与软数据(例如，间隔型数据， 概率分布)。在考虑它们的空间模式的同时合并这些不同的数据层， 因此，成分性质(分类属性的情况)和局部不确定性是具有挑战性的。 该SBIR项目正在开发第一个提供软指标编码和非指标编码工具的商业软件 不确定性的参数地质统计学建模。这款研究产品将是一个独立的桌面空间- 时间(ST)分析和可视化工具，基于BioMedware开发的遗留核心软件。 这些工具将适合于分析健康科学以外的数据，例如在遥感方面， 地球化学、城市基础设施或土壤科学，显著拓宽了最终的商业市场 产品。该项目将实现四个目标： 开发了一种指示克里格法来替代泊松和二项克立格法，用于过滤由 小数问题和分解面率数据(面到点克里金法)，同时避免 产生负克立格法估计。 实现了用于预测分类数据的出现概率的单纯指示克里金法， 使用密歇根州弗林特的服务线路(SL)组合的案例，比较这一点的准确性 构成方法：1)传统的指标克立格法，可能导致负概率 出现次数和概率不等于1，以及2)机器学习和概率的组合 美国SL成分预测领先者BlueConduit使用的贝叶斯数据分析。 开发和测试一个原型模块，该模块将通过软指标编码指导非专家 信息和变异函数建模，然后进行空间内插和交叉验证 BioMedware的时空可视化和分析技术。 进行可用性和用户体验研究，并确定要考虑的其他方法和工具 第二阶段。 这些技术、科学和商业创新将增强我们对地统计学建模的能力 多变量时空现象和计算估计以及尺度上的相关不确定性(例如 地点位置、人口普查区域级别)与环境流行病学最相关。

7. Project Summary/Abstract A key component in any investigation of association and/or cause-effect relationships between the environment and health outcomes is the availability of accurate and precise models of exposure. Because the cost of collecting field data is often prohibitive, it is critical to incorporate any source of secondary information available to supplement sparse datasets. Secondary data can take many forms (e.g., continuous or categorical measurement scale), and display different levels of reliability: hard vs soft data (e.g., interval-type data, probability distributions). Merging these different data layers while accounting for their spatial patterns, compositional nature (case of categorical attributes) and local uncertainty is thus challenging. This SBIR project is developing the first commercial software to offer tools for soft indicator coding and non- parametric geostatistical modeling of uncertainty. The research product will be a stand-alone desktop space- time (ST) analysis and visualization tool, building on the legacy core software developed by BioMedware. These tools will be suited for the analysis of data outside health sciences, such as in remote sensing, geochemistry, urban infrastructure or soil science, broadening significantly the commercial market for the end product. This project will accomplish four aims:  Develop an indicator kriging alternative to Poisson and binomial kriging for filtering noise caused by the small number problem and to disaggregate areal rate data (Area-to-Point kriging), while avoiding the generation of negative kriging estimates.  Implement simplicial indicator kriging for predicting the probability of occurrence of categorical data and, using the case of the composition of service lines (SL) in Flint Michigan, compare the accuracy of this compositional approach to: 1) traditional indicator kriging that can result in negative probabilities of occurrence and probabilities that do not sum to one, and 2) a combination of machine learning and Bayesian data analysis used by BlueConduit, a US leader in SL composition prediction.  Develop and test a prototype module that will guide non-expert through the soft indicator coding of information and variogram modeling, followed by the spatial interpolation and cross-validation based on BioMedware’s space-time visualization and analysis technology.  Conduct a usability and user experience study and identify additional methods and tools to consider in Phase II. These technologic, scientific and commercial innovations will enhance our ability to model geostatistically multivariate space-time phenomena and compute estimates and the associated uncertainty at the scale (e.g. point location, census-tract level) the most relevant for environmental epidemiology.