Oversampling techniques for diabetes classification: A comparative study

Supplementary Material: In this study, several tests were conducted to evaluate the occurrence of overfitting when applying different oversampling variants and machine learning algorithms. Therefore, we have compared the results obtained in the test dataset with the results obtained in the training dataset. Table S[1 - 3] show the F1-Score results obtained for each oversampling variant. The results concerning the application of different machine learning algorithms to training data are presented for each fold (K=10) dataset, as well as the mean of these values.