Table 1. Performance of different SVM classifiers in LOO CV.
Model | C | γ | Th | SN | SP | Accuracy | MCC |
AAC | 19 | 0.001 | −0.9 | 80.00 | 84.45 | 83.65 | 0.557 |
DPC | 2 | 0.01 | −0.7 | 80.00 | 84.45 | 83.65 | 0.557 |
CC | 0 | 4 | −1.0 | 64.00 | 75.65 | 73.55 | 0.328 |
HC | 0 | 0.1 | −1.0 | 68.00 | 74.48 | 73.31 | 0.346 |
MPC | 2 | 0.001 | −1.0 | 77.33 | 76.53 | 76.68 | 0.439 |
ACM | 0 | 0.00001 | −1.0 | 80.00 | 78.29 | 78.60 | 0.479 |
ACHM | 20 | 0.001 | −0.8 | 82.66 | 86.80 | 86.05 | 0.610 |
PSSM-a | 22 | 4 | −0.6 | 80.00 | 87.68 | 86.29 | 0.604 |
PSSM-b | 60 | 0.0001 | −0.6 | 82.66 | 81.81 | 81.97 | 0.541 |
Th–Threshold, SN–sensitivity, SP–specificity, MCC–Matthews Correlation Coefficient.