Towards the Implementation of Eco-epidemiological models of Dengue in Colombia using Machine Learning and Satellite Images

Abstract

Machine learning (ML) presents countless opportunities for population and public health research, and infectious disease modeling is among those. Dengue is a climate-sensitive disease, and, over the last 50 years, its incidence has increased 30-fold, with a distinctive high burden in countries like Colombia. ML and using deep learning on satellite images have gained more attention in recent years due to the amount of heterogeneous data that could inform dengue disease modeling. We introduced a project that aims to build responsible and explainable ML-based dengue models that supports later deployment and implementation. It includes a global health data science approach in Colombia, with the development of open databases, a spatial model for disease mapping, political incidence, and multi-stakeholder collaboration.