Mining Spatiotemporal Diffusion Network: A New Framework of Active Surveillance Planning

Abstract

Infectious diseases pose a constant and serious threat to human life. One way to prevent infectious disease spread is through active surveillance: monitoring patients to discover disease incidences before they get out of hand. However, active surveillance can be difficult to implement, especially when the monitored area is vast and resources are limited. Incidences of infectious disease that arrive with visitors from abroad are a further challenge. When faced with imported incidences and a large region to monitor, it is critical that public health authorities precisely allocate their sparse resources to high-priority areas to maximize the efficacy of active surveillance. In this paper, the difficulties of active surveillance are considered, and we offer a computational framework to address these challenges by modeling and mining the spatiotemporal patterns of infectious risks from heterogeneous data sources. Malaria is used as an empirical case study (with real-world data) to validate our proposed method and enhance our findings.