Hybrid Modeling of the Marburg Outbreak Dynamics in Uíge (2004–2005): Integration of Differential Equations and Machine Learning in a High-Lethality Context

Objective: To develop and apply a hybrid model based on Ordinary Differential Equations (ODEs) and Machine Learning (ML) algorithms for the understanding and control of the Marburg outbreak that occurred in Uíge Province, Angola. Methods: A modified SEIRO model was used with epidemiological data from the WHO for the years 2004-2005. The choice of this empirical data was because this outbreak occurred during this period and only in the province of Uíge. Subsequently, a Random Forest-based regression algorithm was coupled to predict the evolution of the outbreak from environmental and social variables. Simulations were performed in R. Results: The hybrid model accurately identified the points of greatest risk of spread and assessed the impact of interventions such as quarantine and contact tracing. The machine learning model showed a predictive accuracy of over 93%. Conclusion: The hybrid approach proved effective in modeling and predicting epidemic outbreaks, suggesting its application in other infectious contexts.