Class imbalance in datasets is a common issue across various fields, including banking, medicine, and information security. Data augmentation is a frequently used approach to address this problem by generating additional samples of the minority class to rebalance the dataset. Other studies have employed methods such as Generative Adversarial Networks (GAN) and Synthetic Minority Over-sampling Technique (SMOTE)for this purpose. Therefore, this study aims to compare the differences between the two oversampling techniques, GAN and SMOTE, in handling class imbalance problems. The results of this study show the accuracy of distinguishing between real and generated data to determine which method offers a greater advantage. The method demonstrates better performance in multi-class classification tasks. The GAN model can be effectively applied to both binary classification and the generation of diverse samples from minority and majority classes, even in extreme cases where the number of minority samples is tiny. Moreover, in terms of classification accuracy and the quality of generated samples, GAN outperforms SMOTE in data augmentation and oversampling. It maintains strong performance even when the number of instances in the minority class is limited.
Class imbalance in medical classification can substantially affect model performance, particularly when resampling intensity and decision threshold are considered separately. This study systematically investigates their joint effects using the Pima Indians Diabetes, Heart Failure, and Thoracic Surgery datasets. Five classifiers were evaluated with Baseline, SMOTE, and ADASYN under multiple oversampling intensities and decision thresholds ranging from 0.30 to 0.70 using leakage-free, repeated, stratified 10-fold cross-validation with five repetitions. The best mean F1-scores were 0.6890 for PIMA, 0.7684 for Heart Failure, and 0.3375 for Thoracic Surgery, with the preferred threshold and oversampling intensity differing across datasets. A mixed-effects analysis confirmed a significant interaction between oversampling intensity and decision threshold (χ2(16) = 141.77, p < 0.001). In contrast, the matched difference between SMOTE and ADASYN was negligible (ΔF1 = 0.0009, p = 0.663). Thresholds below 0.50 were particularly beneficial for PIMA and Thoracic Surgery, whereas Heart Failure performed best near 0.50. Oversampling also tended to increase Brier scores, indicating that improvements in threshold-specific F1-score did not necessarily correspond to better probability calibration. These findings demonstrate that dataset-specific joint evaluation of oversampling intensity and decision threshold is more informative than selecting a resampling method in isolation.