Influence of Class Imbalance on the Quality of Hydrocracking Unit Failure Prediction Models
Table 3: Balanced model’s evaluation metrics
| F-score for | Decision tree | Random Forest | Logistic regression | Gaussian Bayesian |
| Balanced | 0.569 | 0.601 | 0.513 | - |
| Downsampled | 0.575 | 0.605 | 0.475 | 0.501 |
| Upsampled | 0.569 | 0.635 | 0.478 | 0.507 |
| Upsampled AUC-ROC | 0.824 | 0.852 | 0.729 | 0.755 |