Table 5.
Models results according to different data balancing techniques
| Overall metrics | ||||||
|---|---|---|---|---|---|---|
| Classifier | Data balancing | Acc. (%) | Sen. (%) | Spec. (%) | MF1 (%) | WF1 (%) |
| KNN | — | 72.5 | 84.3 | 51.3 | 68.1 | 71.1 |
| DT | Adasyn | 66.3 | 56.2 | 85.5 | 66.1 | 66.8 |
| SVM | SMOTE | 71.6 | 71.7 | 73.4 | 70.4 | 72.2 |
| Random Forest | SMOTE | 78.3 | 82.4 | 72.1 | 76.5 | 78.4 |
| Majority Voting | — | 72.3 | 79.9 | 58.1 | 69.2 | 72.1 |
| Stacking | SMOTE | 79.8 | 90.2 | 49.8 | 78.0 | 79.9 |
| Bagging | — | 79.8 | 83.6 | 73.6 | 78.0 | 79.9 |
| AdaBoost | Weighted class | 81.4 | 86.5 | 73.8 | 79.5 | 81.4 |
| GBoost | SMOTE | 78.8 | 81.1 | 75.4 | 77.1 | 78.9 |
Values in bold refer to statically significant values.