Table 5 Comparing the results of CBR and DT RBR on WISDM smartphone and smartwatch activity and biometrics dataset before and after SMOTE
From: Building an enhanced case-based reasoning and rule-based systems for medical diagnosis
Approach | Accuracy | Precision | Recall | F1-Score |
|---|---|---|---|---|
Dice coefficient (before SMOTE) | 54.19% | 50% | 46% | 45% |
Jaccard similarity (before SMOTE) | 54.19% | 50% | 46% | 45% |
Logistic regression similarity (before SMOTE) | 73.21% | 55% | 49% | 47% |
Random forest similarity (before SMOTE) | 89.48% | 88% | 85% | 86% |
GaussianNB similarity (before SMOTE) | 75.83% | 70% | 69% | 68% |
DT RBRS (before SMOTE) | 86% | 84% | 83% | 84% |
Dice coefficient (after SMOTE) | 65.54% | 57% | 50% | 53% |
Jaccard similarity (after SMOTE) | 61.23% | 65.17% | 49.41% | 51% |
Logistic regression similarity (after SMOTE) | 64.00% | 61% | 54% | 57% |
Random forest similarity (after SMOTE) | 94.19% | 94% | 93% | 93% |
GaussianNB similarity (after SMOTE) | 70.98% | 69% | 66% | 67% |
DT RBRS (after SMOTE) | 89% | 87% | 88% | 86% |