From: A hybrid CNN-Bi-LSTM model with feature fusion for accurate epilepsy seizure detection
Classifier | Classes | Accuracy | Sensitivity | Specificity | Precision | F1-score |
---|---|---|---|---|---|---|
SVM | C-E | 98.18 | 99.23 | 97.04 | 97.10 | 98.11 |
D-E | 97.50 | 98.84 | 96.40 | 96.09 | 97.39 | |
CD-E | 98.22 | 98.95 | 97.28 | 98.82 | 98.72 | |
Interictal-Ictal | 100.00 | 100.00 | 100.00 | 100.00 | 100.00 | |
Preictal-Ictal | 96.17 | 95.84 | 97.00 | 96.40 | 95.95 | |
CNN | C-E | 99.67 | 99.82 | 99.85 | 99.82 | 99.66 |
D-E | 99.00 | 99.20 | 98.77 | 98.89 | 99.04 | |
CD-E | 98.72 | 98.96 | 98.39 | 99.18 | 99.05 | |
Interictal-Ictal | 100.00 | 100.00 | 100.00 | 100.00 | 100.00 | |
Preictal-Ictal | 95.33 | 94.37 | 96.77 | 96.66 | 95.30 | |
LSTM | C-E | 99.17 | 99.19 | 99.16 | 99.18 | 99.17 |
D-E | 100.00 | 100.00 | 100.00 | 100.00 | 100.00 | |
CD-E | 99.06 | 99.27 | 98.78 | 99.30 | 99.28 | |
Interictal-Ictal | 100.00 | 100.00 | 100.00 | 100.00 | 100.00 | |
Preictal-Ictal | 94.17 | 96.27 | 92.25 | 92.85 | 94.34 | |
Bi-LSTM | C-E | 99.33 | 99.29 | 99.33 | 99.35 | 99.31 |
D-E | 100.00 | 100.00 | 100.00 | 100.00 | 100.00 | |
CD-E | 99.10 | 99.68 | 97.95 | 99.01 | 99.34 | |
Interictal-Ictal | 100.00 | 100.00 | 100.00 | 100.00 | 100.00 | |
Preictal-Ictal | 97.80 | 98.39 | 96.55 | 98.39 | 99.20 | |
CNN-Bi-LSTM | C-E | 99.75 | 99.69 | 99.84 | 99.83 | 99.74 |
D-E | 100.00 | 100.00 | 100.00 | 100.00 | 100.00 | |
CD-E | 99.11 | 99.48 | 98.45 | 99.14 | 99.31 | |
Interictal-Ictal | 100.00 | 100.00 | 100.00 | 100.00 | 100.00 | |
Preictal-Ictal | 98.83 | 99.33 | 98.46 | 98.27 | 98.76 |