Bayesian Regression Model to Analyze, Predict and Control the Spreading of COVID-19 in Germany with High Spatial Resolution

利用贝叶斯回归模型以高空间分辨率分析、预测和控制德国 COVID-19 的传播

• 批准号: 492350939
• 负责人: Professor Dr. Gordon Pipa
• 金额: --
• 依托单位: Institut für Kognitionswissenschaft
• 依托单位国家: 德国
• 项目类别: Research Grants
• 财政年份: 2021
• 资助国家: 德国
• 起止时间: 2020-12-31 至 2022-12-31
• 项目状态: 已结题
• 来源: https://gepris.dfg.de/gepris/projekt/492350939?language=en
• 关键词: Bayesian; Regression; Model; Analyze; Predict

Aim of the proposal is to provide a fully data-driven analysis and forecast of spatial temporal dynamics, with an improved spatial resolution that will enable decision-makers to judge the current and predicted dynamics, to assess the reliability and possible variations of the predictions, to plan and adjust regulation to control the outbreak. The model will include explicit factors that model the spreading across regions that will allow visualizing the temporal spreading and the identification of driving factors of the disease. To this end, a fully datadriven Bayesian regression model will be used on two spatial scales. Firstly, on the level of counties (Landkreise), and secondly on the level of regions (~40) inside the county of Osnabrück. In the second phase of the project, Oldenburg will be included as a second high resolution spatial model. The models will be an extension of the established fully Bayesian regression model, initially developed in cooperation with the RKI, and later adapted to COVID-19 (https://covid19-bayesian.fz-juelich.de/), and that runs prediction on the level of counties since August 2020, and operates as a prototype on the finer level since January 2021.

该提案的目的是对时空动态提供完全由数据驱动的分析和预测，提高空间分辨率，使决策者能够判断当前和预测的动态，评估预测的可靠性和可能的变化，规划和调整监管以控制疫情。该模型将包括对跨区域传播进行建模的明确因素，这将允许可视化时间传播并识别疾病的驱动因素。为此，将在两个空间尺度上使用完全由数据驱动的贝叶斯回归模型。首先，在县（Landkreise）一级，其次是在奥斯纳布吕克县内的地区（~40）一级。在该项目的第二阶段，奥尔登堡将作为第二个高分辨率空间模型。该等模型将是已建立的全贝叶斯回归模型的扩展，该模型最初与RKI合作开发，后来适应COVID-19（https：//covid19-bayesian.fz-juelich.de/），自二零二零年八月起在县一级进行预测，并自二零二一年一月起在更精细的层面上作为原型运行。