Predicting and Monitoring of Preeclampsia: An Evaluation with Machine Learning Techniques

Preeclampsia (PE) emerges following the first 20 weeks of pregnancy, marked by the presence of proteinuria and hypertension. This condition is linked to placental dysfunction and hypertension induced due to pregnancy. It can impact both maternal health and fetal growth, potentially leading to low birth weights - a factor contributing to the risk of neonatal mortality. Hypertensive disorders during pregnancy impact about 10% of global pregnancies. Within this examination, we delve into the involvement of Machine Learning (ML) techniques in forecasting PE. This review seeks to offer a brief summary of recent advancements in predicting, diagnosing, and monitoring PE through ML algorithms by examining the potential implications through stemming from the capability of various ML algorithms (RF, DT, LR, SVM). In the proposed model 14 factors were included, such as maternal traits, health history, prenatal test results, and ultrasound findings. Cross-validation is used to assess the differentiation, and area under the receiver operating curve (AUROC) is observed. Hence, the ML techniques (SVM, LR, DT and RF) were executed, and patient-centric healthcare information on PE can be effectively retrieved using Random Forest (RF).