Table 1 Performance of the SVM using different feature subsets selected by ReliefF
Feature subset | Sensitivity (%) | Specificity (%) | MCC | Test Accuracy (%) | CV Accuracy |
|---|---|---|---|---|---|
75 features | 69.50 | 77.13 | 0.4663 | 73.86 | 77.16 |
100 features | 74.47 | 84.21 | 0.5904 | 80.06 | 80.00 |
200 features | 68.09 | 81.58 | 0.5022 | 75.83 | 78.00 |
300 features | 67.38 | 82.11 | 0.5017 | 75.83 | 78.67 |
400 features | 64.54 | 86.32 | 0.5260 | 77.04 | 78.00 |
500 features | 65.96 | 85.79 | 0.5323 | 77.34 | 78.00 |
All features | 63.12 | 78.19 | 0.4182 | 71.73 | 75.16 |