Table 4 Performance of the SVM classifier using different resampling techniques, reported as mean ± standard deviation across all datasets. The results include evaluation metrics such as AUC, F1-score, G-mean, Recall, and Precision. The proposed Borderline-Shifting Oversampling (BSO) method demonstrates superior or competitive performance compared with existing resampling approaches, highlighting its effectiveness in addressing class imbalance while maintaining classifier stability.
From: An approach for handling imbalanced datasets using borderline shifting
Technique | F1-score | G-mean | Precision | Recall | AUC |
|---|---|---|---|---|---|
Without | 0.93 ± 0.02 | 0.54 ± 0.08 | 0.52 ± 0.14 | 0.54 ± 0.08 | 0.81 ± 0.18 |
NearMiss | 0.55 ± 0.14 | 0.55 ± 0.11 | 0.55 ± 0.14 | 0.55 ± 0.11 | 0.57 ± 0.17 |
Random Undersampling | 0.72 ± 0.17 | 0.76 ± 0.11 | 0.77 ± 0.14 | 0.76 ± 0.11 | 0.82 ± 0.17 |
Random Oversampling | 0.79 ± 0.11 | 0.82 ± 0.08 | 0.82 ± 0.13 | 0.82 ± 0.08 | 0.88 ± 0.07 |
SMOTE | 0.81 ± 0.10 | 0.84 ± 0.08 | 0.84 ± 0.08 | 0.84 ± 0.08 | 0.90 ± 0.07 |
Borderline SMOTE | 0.82 ± 0.11 | 0.84 ± 0.08 | 0.83 ± 0.08 | 0.84 ± 0.08 | 0.90 ± 0.07 |
SMOTE Tomek | 0.80 ± 0.11 | 0.83 ± 0.10 | 0.84 ± 0.13 | 0.83 ± 0.07 | 0.89 ± 0.06 |
SMOTEENN | 0.86 ± 0.18 | 0.89 ± 0.09 | 0.89 ± 0.13 | 0.89 ± 0.09 | 0.92 ± 0.07 |
Our Approach | 0.91 ± 0.03 | 0.91 ± 0.04 | 0.91 ± 0.03 | 0.91 ± 0.04 | 0.95 ± 0.07 |
