Automated Machine Learning-Based System for the Prediction of Maternal Health Indicators and High-Risk Pregnancy

Maternal health is critical aspect affecting the wellbeing of both mother and child. Despite many advancements in the field of healthcare, many women still face pregnancy health complications due to factors such as socioeconomic factors and pre-existing medical conditions. Identifying risk in pregnancies remains challenging due to these risk factors. In this study, we have utilized machine learning algorithms to identify high risk pregnancy. Using comprehensive dataset of medical records of pregnant women, we applied tree based models, logistic regression, support vector, ensemble methods including stacking, voting classifier, KNN and naïve bayes algorithm for classifying high risk and non-risk pregnancies. The results demonstrate that ensemble methods such as stacking classifier outperform traditional machine learning models with predictive accuracy of 98%, sensitivity of 0.97, precision of 0.97, recall of 0.97, providing a robust framework for early pregnancy risk detection. This study highlights the potential of computational methods in enhancing maternal healthcare and mitigating risks during pregnancy.