Table 1 Groupwise average accuracies of various feature extraction methods combined with various classifiers applied to classify sets A, D, and E (standard deviations are noted in parentheses).
From: Combination of EEG Complexity and Spectral Analysis for Epilepsy Diagnosis and Seizure Detection
Classifier | Feature selection | A | D | E | Accuracy |
|---|---|---|---|---|---|
LLS | All features | 100.00 (0.0) | 95.00 (3.9) | 95.50 (1.6) | 96.83 (1.2) |
| Â | GA | 99.25 (1.2) | 94.50 (3.3) | 94.75 (3.6) | 96.17 (1.9) |
| Â | ApEn + AR model | 98.50 (2.4) | 95.50 (2.6) | 95.50 (3.1) | 96.50 (1.3) |
LDA | All features | 100.00 (0.0) | 94.50 (3.9) | 95.75 (1.7) | 96.75 (1.1) |
| Â | GA | 99.50 (1.1) | 95.00 (4.4) | 94.50 (2.3) | 96.33 (1.9) |
| Â | ApEn + AR model | 97.50 (2.6) | 96.00 (2.9) | 95.75 (3.1) | 96.41 (1.6) |
BP | All features | 98.75 (3.1) | 96.50 (2.9) | 97.00 (2.0) | 97.42 (1.4) |
| Â | PCA | 100.0 (0.0) | 97.75 (1.8) | 97.00 (2.6) | 98.25 (1.5) |
| Â | GA | 98.50 (1.7) | 91.50 (5.2) | 97.50 (2.4) | 95.83 (2.0) |
| Â | ApEn + AR model | 99.25 (1.2) | 93.00 (5.5) | 89.50 (6.0) | 93.92 (3.0) |
LISVM | All features | 99.50 (1.1) | 97.00 (2.6) | 98.25 (1.2) | 98.25 (1.1) |
| Â | PCA | 99.75 (0.8) | 98.00 (2.0) | 97.25 (1.4) | 98.33 (0.6) |
| Â | GA | 98.75 (2.1) | 93.50 (5.0) | 98.00 (1.1) | 96.75 (1.7) |
| Â | ApEn + AR model | 99.75 (0.8) | 94.25 (3.1) | 94.50 (5.0) | 96.17 (2.2) |
RBFSVM | All features | 99.75 (0.8) | 97.75 (1.8) | 97.75 (1.4) | 98.42 (0.8) |
| Â | PCA | 99.75 (0.8) | 98.25 (1.8) | 98.00 (1.6) | 98.67 (0.7) |
| Â | GA | 99.50 (1.6) | 95.00 (3.7) | 96.50 (2.7) | 97.00 (1.7) |
| Â | ApEn + AR model | 99.75 (0.8) | 92.25 (3.2) | 91.75 (3.9) | 94.58 (1.8) |