African Journal of Health Informatics

Machine Learning Approaches for Early Detection of Malaria in Sub-Saharan Africa

Emeka Okafor, Amina Yusuf, Chukwuemeka Nwachukwu

This study investigates the application of supervised machine learning algorithms — including Random Forest, Support Vector Machines, and Gradient Boosting — to the early diagnosis of malaria using clinical and haematological data collected from tertiary hospitals across Sub-Saharan Africa. A dataset of 14,200 patient records was used for training and evaluation. The proposed ensemble model achieved a sensitivity of 94.3% and specificity of 91.7%, outperforming conventional microscopy-based screening in resource-constrained settings. The findings suggest that low-cost, deployable ML pipelines can significantly reduce diagnostic delays and improve patient outcomes in endemic regions.