Table 10. Classification results of various methods in terms of classification accuracy, sensitivity, specificity, number of selected features and support vectors, and model parameters for Parkinson’s disease dataset (PDD).
| Criteria | Methods | ||||
|---|---|---|---|---|---|
| GAFS | PSOFS | TVPSOFS | FOASVM | CIFOA-SVM | |
| Sensitivity | 0.9890 + 0.0017 | 0.9703 + 0.0019 | 0.6533 + 0.0016 | 0.9790 + 0.0029 | 0.9810 + 0.0013 |
| Accuracy | 0.9621 + 0.0022 | 0.9157 + 0.0024 | 0.9473 + 0.0011 | 0.9615 + 0.0017 | 0.9631 + 0.0006 |
| Specificity | 0.9352 + 0.0069 | 0.8333 + 0.0187 | 0.8666 + 0.0433 | 0.9440 + 0.0025 | 0.9452 + 0.0027 |
| N of features | 10.2 + 6.92 | 15.3 + 3.61 | 12.5 + 8.45 | 22.0 + 0.0 | 12.1 + 1.59 |
| N of SVs | 78.1 | 82.6 | 81.2 | 129.8 | 75.4 |
| Parameter C | 594.1864 | 449.5192 | 562.3227 | 543.8755 | 472.205 |
| Parameter γ | 0.4939 | 0.5348 | 0.3189 | 0.3495 | 0.4518 |