Predicting Maternal Health Risks in Sierra Leone Using Machine Learning Models

Maternal health risk prediction can support the early identification of pregnancies that may require closer clinical attention, particularly in resource-constrained settings. This study evaluates seven supervised machine-learning classifiers— Logistic Regression, Support Vector Machine (SVM) with Polynomial and Radial Basis Function (RBF) kernels, Gradient Boosting Machine (GBM), XGBoost, LightGBM, and CatBoost—for three-class maternal health-risk classification. The primary analysis uses a 500-record dataset obtained from Poly Clinic, Peninsula Road, Sierra Leone, containing eight predictors: age, systolic blood pressure, diastolic blood pressure, blood glucose, BMI, heart rate, parity, and hemoglobin.