Predicting Newborn Low Birth Weight: A Machine Learning Approach Using Maternal Health and Demographic Data

Low birth weight (LBW), defined by the World Health Organization as a birth weight under 2500 grams, remains a major global health challenge due to its association with increased risks for adverse neonatal and long-term health outcomes. This study aims to identify maternal health factors that contribute to LBW and to develop predictive models to support early identification of at-risk pregnancies. Two machine learning models, Logistic Regression (LR) and Random Forest (RF), were developed to analyze the relationship between maternal factors and LBW outcomes. The RF model achieved an accuracy of 96.36%, demonstrating robust predictive performance across metrics, making it effective for applications prioritizing accurate classification. Conversely, the LR model, with an accuracy of 91.95%, offered valuable interpretability, identifying significant predictors of LBW, including gestational age, maternal weight gain during pregnancy, anemia, and chronic hypertension. The consistency of key predictors across both models reinforces the importance of these maternal health factors in LBW outcomes. The findings suggest that predictive models can effectively assist healthcare providers in identifying high-risk pregnancies and allocating resources for early interventions. Practical recommendations include enhanced prenatal care for at-risk groups, public health education on maternal health, and policies supporting accessible maternal care.