Machine Learning Applications in Malaria Elimination Programs: Comparing Vector Control Strategies Across West Africa and Former Endemic Regions in the Southern United States

The application of machine learning (ML) technologies in malaria elimination programs represents a paradigm shift in vector-borne disease control strategies. This review examines the comparative implementation of ML-based approaches in vector control programs across West Africa and historically endemic regions in the Southern United States. Through systematic analysis of recent literature, we evaluate the effectiveness of ML algorithms including support vector machines, random forests, deep learning models, and predictive analytics in malaria vector surveillance and control. Our findings reveal that while West African programs leverage ML primarily for outbreak prediction and vector habitat mapping using drone imagery and environmental data, the historical elimination success in the Southern United States provides valuable lessons for contemporary ML-enhanced programs. The review demonstrates that machine learning models such as support vector machines, decision trees, random forests, Extreme Gradient Boosting, logistic regression, K-Nearest Neighbours, Naïve Bayes, and multilayer perceptron have been greatly used to predict malaria using socioeconomic and environmental variables. Current applications show promise in drone imagery and deep learning analysis for targeted vector surveillance, enabling more precise identification of mosquito breeding sites. This comparative analysis highlights the evolution from traditional vector control methods to sophisticated ML-driven approaches, offering insights for optimizing future malaria elimination strategies in endemic regions.