Exploring Maternal Health Indicators as Predictors for Low Birth Weight in Neonatal: A Data-Driven Machine Learning Approach

Abstract

Low birth weight (LBW) remains a significant public health concern, contributing to neonatal morbidity and mortality. Identifying key maternal health indicators that predict LBW can facilitate early intervention and improved healthcare strategies. This research paper explores the potential of various maternal health factors as predictors for LBW using a data-driven machine learning approach. We analyze a dataset comprising maternal age, weight, smoking status, hypertension, race, and prenatal care factors, alongside the birth weight of newborns. Using logistic regression and several machine learning algorithms, including decision trees, random forests, and support vector machines, we evaluate the performance of these models in classifying newborns as having low or normal birth weight. The study highlights the importance of maternal smoking, age, and weight as primary predictors of LBW, while also assessing the significance of factors such as hypertension and the number of prenatal visits. The results indicate that machine learning models, particularly random forests, can provide high accuracy in predicting LBW outcomes, thereby offering valuable insights for healthcare providers in managing pregnancies at risk. The findings underscore the need for targeted interventions focused on maternal health to reduce the incidence of low birth weight, with implications for prenatal care policies and maternal health programs.\