Table 12 Influence of SMOTE over the classifiers.
From: Mitigating class imbalance in churn prediction with ensemble methods and SMOTE
Model | Accuracy | Balanced Accuracy | Weighted Precision | Weighted Recall | Weighted Recall |
---|---|---|---|---|---|
LR | 0.811 | 0.593 | 0.669 | 0.373 | 0.442 |
LR + SMOTE | 0.801 | 0.650 | 0.788 | 0.801 | 0.793 |
GB | 0.866 | 0.723 | 0.813 | 0.577 | 0.663 |
GB + SMOTE | 0.850 | 0.758 | 0.849 | 0.850 | 0.849 |
XGB | 0.858 | 0.724 | 0.763 | 0.589 | 0.657 |
XGB + SMOTE | 0.855 | 0.755 | 0.850 | 0.855 | 0.755 |
LGBM | 0.860 | 0.726 | 0.774 | 0.590 | 0.662 |
LGBM + SMOTE | 0.804 | 0.755 | 0.646 | 0.804 | 0.716 |
ADA | 0.859 | 0.717 | 0.781 | 0.572 | 0.650 |
ADA + SMOTE | 0.845 | 0.771 | 0.850 | 0.845 | 0.847 |