Table 6 Sample classification, thresholding approach results (expressed in percentage), RFE feature selection
| Â | Feature Selection | ||||
|---|---|---|---|---|---|
Threshold | Classifier | Accuracy | Sensitivity | Specificity | F score |
50% | MLP | 76.6 ±42.57 | 64.70 | 90.69 | 76.37 |
45% | MLP | 77.66 ±41.88 | 68.62 | 88.37 | 77.58 |
40% | AdaBoost | 79.79 ±40.37 | 70.59 | 90.70 | 79.71 |
35% | AdaBoost | 80.85 ±39.56 | 74.51 | 88.37 | 80.85 |
35% | SVM | 78.72 ±41.15 | 70.58 | 88.37 | 78.67 |
30% | AdaBoost | 80.85 ±39.56 | 78.43 | 83.72 | 80.89 |
30% | SVM | 78.72 ±41.15 | 72.55 | 86.05 | 78.72 |
25% | AdaBoost | 84.04 ±36.82 | 86.27 | 81.39 | 84.03 |
25% | RF | 78.72 ±41.15 | 76.47 | 81.39 | 78.76 |
20% | AdaBoost | 80.85 ±39.56 | 88.24 | 72.09 | 80.66 |
15% | RF | 77.66 ±41.88 | 82.35 | 72.09 | 77.57 |
10% | RF | 75.53 ±43.22 | 86.27 | 62.79 | 75.09 |
10% | SVM | 74.47 ±43.84 | 88.24 | 58.14 | 73.74 |
5% | RF | 73.4 ±44.42 | 92.15 | 51.16 | 72.03 |