Table 7 Comparison of optimization techniques.
Dataset | Optimizer | Training loss | Validation loss | Accuracy (%) | F1 Score (%) | ROC-AUC Score (%) | Convergence epochs | Training stability |
|---|---|---|---|---|---|---|---|---|
Faisalabad | Adam | 0.1 | 0.2 | 99.51 | 99.5 | 99.35 | 14 | High |
SGD | 0.3 | 0.4 | 85 | 80 | 87 | 32 | Moderate | |
RMSprop | 0.25 | 0.35 | 88 | 85 | 90 | 21 | High | |
Adagrad | 0.28 | 0.38 | 86 | 82 | 90 | 25 | Moderate | |
CVD | Adam | 0.1 | 0.25 | 98.76 | 99 | 99.34 | 15 | High |
SGD | 0.35 | 0.4 | 87 | 85 | 89 | 36 | Low | |
RMSprop | 0.39 | 0.45 | 85 | 89 | 90 | 23 | Moderate | |
Adagrad | 0.3 | 0.35 | 89 | 85 | 85 | 27 | Moderate | |
Heart failure | Adam | 0.15 | 0.2 | 99.07 | 99.22 | 99.43 | 13 | High |
SGD | 0.2 | 0.35 | 89 | 85 | 87 | 34 | Moderate | |
RMSprop | 0.3 | 0.45 | 85 | 87 | 85 | 22 | Moderate | |
Adagrad | 0.25 | 0.4 | 80 | 80 | 90 | 26 | Moderate |
