Table 3.
Sensitivity and specificity of different features and optimizers by 10-fold cross-validation.
| Optimizer | Time-domain feature | FC feature | Wavelet feature | |||
|---|---|---|---|---|---|---|
| Sensitivity | Specificity | Sensitivity | Specificity | Sensitivity | Specificity | |
| GA | 87.40 ± 1.43% | 82.10 ± 2.96% | 80.80 ± 2.04% | 83.80 ± 1.69% | 91.70 ± 1.70% | 77.30 ± 2.06% |
| (91.00%) | (82.00%) | (84.00%) | (85.00%) | (94.00%) | (80.00%) | |
| PSO | 85.40 ± 2.01% | 82.10 ± 2.77% | 78.50 ± 2.46% | 80.30 ± 3.33% | 90.40 ± 1.51% | 73.60 ± 2.37% |
| (87.00%) | (85.00%) | (80.00%) | (83.00%) | (91.00%) | (78.00%) | |
| pGAPSO-I | 86.20 ± 1.40% | 83.60 ± 2.55% | 79.00 ± 2.94% | 83.50 ± 2.32% | 88.90 ± 2.08% | 75.80 ± 1.99% |
| (86.00%) | (87.00%) | (85.00%) | (82.00%) | (90.00%) | (77.00%) | |
| pGAPSO-II | 85.30 ± 1.16% | 84.00 ± 1.76% | 80.10 ± 3.38% | 82.80 ± 3.36% | 89.60 ± 2.22% | 75.40 ± 3.24% |
| (85.00%) | (87.00%) | (80.00%) | (87.00%) | (90.00%) | (79.00%) | |
| pGAPSO-III | 85.70 ± 0.95% | 83.60 ± 2.22% | 78.90 ± 3.14% | 82.60 ± 3.34% | 90.40 ± 1.58% | 73.30 ± 2.11% |
| (87.00%) | (86.00%) | (79.00%) | (89.00%) | (88.00%) | (78.00%) | |
| sPSOGA | 86.20 ± 2.25% | 80.30 ± 2.91% | 78.00 ± 3.56% | 83.50 ± 3.78% | 89.90 ± 1.52% | 73.00 ± 2.16% |
| (90.00%) | (81.00%) | (84.00%) | (86.00%) | (92.00%) | (78.00%) | |
| sGAPSO | 85.90 ± 2.08% | 83.40 ± 2.95% | 81.20 ± 2.20% | 82.40 ± 0.97% | 90.30 ± 1.70% | 76.10 ± 2.96% |
| (86.00%) | (87.00%) | (85.00%) | (83.00%) | (91.00%) | (79.00%) | |
The values in parathesis corresponded to the results using optimal channel combination.