The AFRIDIARRHEA framework combines Bayesian modeling with machine learning to map childhood diarrheal disease burden across Kenya, Zimbabwe, and Somaliland, identifying rotavirus and Shigella as primary drivers of pathogen-attributed mortality. Zimbabwe showed the highest mortality and morbidity burden, while Somaliland had the greatest hospitalization rates. This computational advance addresses a critical gap in African health surveillance where traditional monitoring systems fail to capture the true burden of diarrheal diseases—the second leading cause of death in children under five globally. The framework's ability to integrate environmental data, pathogen attribution, and uncertainty quantification could transform how public health officials allocate scarce resources for vaccination programs and water sanitation interventions. However, this preprint awaits peer review, and the analysis relies on synthetic rather than real-world datasets, limiting immediate applicability. While methodologically sophisticated, the true test will be validation against actual surveillance data. The work represents incremental progress in computational epidemiology but could prove transformative for child health policy in data-scarce African contexts if validated through peer review.