ABSTRACT:
All pregnancies bear risks and it is necessary to care for all expectant mothers during and after pregnancy. However, high-risk pregnancies necessitate more attention from the early stages to prevent and anticipate possible complications. With the recent ease in acquiring and storing large amounts of data digitally, it is now possible to use certain health-related predictors to determine the risk of pregnancy complications before they arise. The ability of machine learning (ML) models to learn complexities in data makes them better suited for this task than conventional statistical methods. In this study, we retrieved data from the University of California, Irvine (UCI) machine learning repository to train machine learning models for pregnancy risk prediction, dividing the pregnancies risks into three classes: low-risk, mid-risk, and high-risk. A total of six ML algorithms were used to make predictions on the dataset, including logistic regression, random forest, support vector classifier, k-nearest neighbours, decision tree, naive Bayes. The random forest model had the best performance, with an area under the ROC curve (AUC) score of 0.7. The use of maternal factors to predict pregnancy risk classes can facilitate pre-labour decisions and better management of pregnancy complications.