The most effective machine learning algorithms in predicting gestational diabetes: A systematic review

Background and Aims: Gestational diabetes mellitus (GDM) is a condition in which the body is unable to effectively utilize insulin, leading to insulin resistance and glucose intolerance. Timely prediction of this disorder can aid in better management of the health of both mother and baby. The purpose of this study is to systematically review published papers on gestational diabetes prediction using machine learning (ML) algorithms and to introduce and compare the most efficient of these algorithms. Methods: A total of ۱۰۶۲ publications were found, of which ۳۴ studies were considered in this review. The authors conducted a systematic search in databases including PubMed, Scopus, and Embase using the keywords 'Artificial Intelligence,' 'Predictive Modeling,' 'Machine Learning,' and 'gestational diabetes' from ۲۰۲۰ to March ۲۰۲۵. The screening, extraction and analysis of the identified articles were performed by following the PRISMA guidelines. Results: A total of thirty-four articles met the inclusion criteria. Parameters for selecting the most efficient machine learning algorithm include accuracy, sensitivity, specificity, and precision criteria. Among them, the Xgboost algorithm was identified as the best and most efficient machine learning algorithm for predicting gestational diabetes in ۳۰% of the articles (n = ۱۰). Other algorithms recognized as the best and most effective methods in these papers included Random Forest (n = ۸), K-Nearest Neighbors (n = ۴), Logistic Regression (n = ۳), Artificial Neural Networks (n = ۳), Gradient Boosting Machine (n = ۴), and Bayesian Classifier (n = ۲). Analysis of variables also showed that age, body mass index, and family history of diabetes are the main influential factors in predicting this disease. Conclusion: Machine learning algorithms are recognized as effective tools in predicting GDM. The findings of this study can serve as a basis for future research aimed at improving

کلیدواژه ها:

Gestational diabetes ، Machine learning algorithms ، Prediction ، Artificial Intelligence ، Predictive Modeling