Predicting Maternal Health Outcomes Using Machine Learning Models

2025/01/27 by OMANKWU, O. C. B., ETUK, ENEFIOK
#Algorithms #Childbirth #Classification #Machine Learning #Maternal Health #Predictive Models

paper · doi:10.60787/jnamp.v68no1.423

Abstract

This study analyzed maternal health data and developed three analytical models to predict the likelihood of adverse maternal health outcomes during childbirth. The models were evaluated and compared for accuracy to identify the factors that influence maternal health and the potential causes of maternal complications. The three models developed in this study are kNearest Neighbors classification, Decision Trees, and Random Forest classifications. The results indicated that the age of the woman, her level of education, occupation, and location are significant factors that could determine maternal health outcomes during childbirth. Furthermore, the study showed that the Random Forest classification algorithm provided superior results for predicting maternal complications compared to the kNearest Neighbors classification and Decision Trees models. The findings demonstrate that with data on a woman's age, education level, occupation, and location, it is possible to predict maternal health outcomes during childbirth.

Related

• Using Interpretable Machine Learning to Predict Maternal and Fetal Outcomes
• Novel machine learning applications in peripartum care: a scoping review
• Predicting Fetal Birthweight from High Dimensional Data using Advanced Machine Learning
• Closing the Gap in High-Risk Pregnancy Care Using Machine Learning and Human-AI Collaboration
• Using Kernel Methods and Model Selection for Prediction of Preterm Birth