Inferring the hidden and long-range dengue transmission routes in Singapore
Machine Learning: Health · 2025
DOI: 10.1088/3049-477x/ae16ab ↗Kurz gefasst
Fachartikel in der Zeitschrift Machine Learning: Health, 2025. Verfasst von Zhen Yuan Yeo und 6 weiteren Autor:innen. Die Arbeit nennt 78 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: NUS Early Career Research Award, NUS Institute of Data Science, NUS Reimagine grant.
Abstract (Originalsprache)
The dengue virus, a major global health problem, is transmitted via a vector-host epidemic model (i.e. mosquito to human). Even though mosquitoes have a limited flight range, dengue spreads much further, suggesting that human movement plays a significant role in its transmission. To model this, we developed a host–host space-time transmission framework by simplifying the traditional host–vector-host model. Using simulated infection cascades based on residential locations and report times, we demonstrated that it is possible to infer hidden transmission routes between infected locations. Applying our method to Singapore’s dengue data from 2008 to 2022, we found that the inferred transmission networks captured informative patterns and shifts annually. While our model could reasonably predict future infections, its precision was limited by the stochastic nature of real-world outbreaks and incomplete case reporting. We improved prediction accuracy by incorporating historical hotspot patterns and virus genetic fingerprinting data. Our main contribution is demonstrating that hidden long-range dengue transmission routes, likely driven by human mobility, can be inferred directly from infection surveillance data at a fine spatial resolution of 1 km ^2 , without requiring explicit human mobility datasets. Our framework allows weekly forecasts of new infections at a 1 km ^2 resolution and can be applied to other diseases with spatial and temporal tagging. The observed shifts of transmission networks highlight the limitations of static mobility-based models. This work shows that surveillance and genomic data can be leveraged to guide dynamic, data-driven disease control strategies.
Förderung
- NUS Early Career Research Award
- NUS Institute of Data Science
- NUS Reimagine grant
Zitieren
APA
Yeo, Z. Y., Hapuarachchi, H. C., Ong, J., Yeo, G., Koo, C., Ng, L. C., & Loh, N. D. (2025). Inferring the hidden and long-range dengue transmission routes in Singapore. Machine Learning: Health. https://doi.org/10.1088/3049-477x/ae16ab
Vancouver
Yeo ZY, Hapuarachchi HC, Ong J, Yeo G, Koo C, Ng LC, et al. Inferring the hidden and long-range dengue transmission routes in Singapore. Machine Learning: Health. 2025. doi:10.1088/3049-477x/ae16ab
BibTeX
@article{Yeo2025,
author = {Yeo, Zhen Yuan and Hapuarachchi, Hapuarachchige Chanditha and Ong, Janet and Yeo, Gladys and Koo, Carmen and Ng, Lee Ching and Loh, N Duane},
title = {Inferring the hidden and long-range dengue transmission routes in Singapore},
journal = {Machine Learning: Health},
year = {2025},
doi = {10.1088/3049-477x/ae16ab}
}
EUVIMED