Machine Learning Models for Predicting the Practice of Exclusive Breastfeeding and Underweight Status Among Under-Five Children

Underweight among children and suboptimal exclusive breastfeeding (EBF) remain two prevalent public health issues in developing and low to middle-income countries like Nigeria. Traditional statistical techniques have limitations in modelling complex relationships among maternal sociodemographic elements that pertain to infant nutrition. This study aimed to propose the application of machine learning (ML) models to predict underweight status and the practice of EBF based on maternal sociodemographic characteristics. This approach aimed to improve the identification of at-risk groups and inform targeted public health interventions. Random forest machine learning models were developed to predict non-exclusive breastfeeding mothers and infant underweight status. The models demonstrated high predictive performance, with testing accuracies of 77% and 88%, respectively. These findings highlight the potential of machine learning tools for early identification of at-risk infants and targeted maternal-child health interventions. Strengthening breastfeeding promotion programs and integrating predictive analytics into healthcare systems may enhance child health outcomes in similar resource-limited settings.

Key Words: machine learning, underweight, exclusive breastfeeding, maternal factors, infant nutrition.