Machine Learning_Based Paradigm for Diagnosis of Gestational Diabetes

One of the issues related to pregnancy is gestational diabetes mellitus (GDM). The risk of exploration in the GDM is lowered with early detection. The study aims to develop a machine learning-based paradigm for the detection of the GDM. Four distinct classifiers-logistic regression, decision tree, gradient boosting, and support vector machine were used. Nine factors of the clinical setting are distinguished in each case. A data set called PIMA for diabetes is used as the data source. The challenge of the existence of missing values is solved by substituting the mean and median values alternatively. Therefore, three scenarios have been applied to the dataset. The performance of the classifiers was assessed by measuring accuracy, sensitivity, specificity, and precision. Results yielded reliable outcomes for the proposed paradigm. In specific, the best results were obtained for the gradient boosting classifier by substituting the missing value by the median value. Results demonstrate the effectiveness of the paradigm.