【114/12/18】15:20-16:10 龔一鴻 助理教授(輔仁大學統計資訊學系)
發佈日期 :
2025-10-31
| Colloquium | |
|
|
|
| Time | 2025-12-18 15:20-16:10 |
| Venue | 數學館1樓3174教室 |
| Speaker | 龔一鴻 助理教授(輔仁大學統計資訊學系) |
| Title | Joint Modeling of Spatial Clustering and Environmental Risk Factors |
| Abstract | Emerging and re-emerging infections—such as influenza, dengue, enterovirus, and COVID-19—continue to challenge global health systems, underscoring the importance of understanding the contextual drivers of disease transmission. Using Taiwan as a natural laboratory, this study quantifies how demographic and socioeconomic structures influence the spatial variation of COVID-19 spread. We propose a mixture-based spatial scan statistic that simultaneously accounts for spatial autocorrelation and covariate effects. This approach allows for the detection of clusters based on established risk factors while also uncovering potential unknown geographic risks. The results reveal that transmission risks are elevated in densely populated areas and communities characterized by higher child dependency ratios and larger proportions of low- to middle-income households. Compared with conventional scan statistics that neglect residual spatial dependence, the proposed method delineates more precise and statistically significant hotspots. Overall, the integration of spatial analytical techniques with contextual risk analysis provides a robust foundation for evidence-based epidemic surveillance and more targeted public health interventions. The approach is generalizable to other infectious diseases influenced by social structure and spatial dependence.
Keywords: Spatial Epidemiology, Scan Statistic, Hotspot Detection, Risk Factors, Spatial Correlation |
|
|
|