Table 5.
Average performance on testing dataset for the different feature subset models.
| Input Features Acronym | F1-Score (%) | Accuracy (%) | Sensitivity (%) | Specificity (%) | PPV (%) | NPV (%) | AUC (%) |
|---|---|---|---|---|---|---|---|
| EnSFS | 63.7 ± 5.1 | 63 ± 4.6 | 65.3 ± 7.6 | 60.6 ± 7 | 62.5 ± 4.6 | 63.8 ± 5.4 | 66.3 ± 4.71 |
| EnALL | 76.8 ± 3.2 | 74.6 ± 4.2 | 83.9 ± 4.8 | 65.4 ± 8.7 | 71.1 ± 5 | 80.4 ± 4.5 | 80.8 ± 4.76 |
| Linear | 87.6 ± 2.2 | 86.3 ± 2.6 | 96.4 ± 3.2 | 76.3 ± 5.4 | 80.4 ± 3.5 | 95.6 ± 3.6 | 90 ± 2.5 |
| LNL | 88.4 ± 2.3 | 87.3 ± 2.7 | 96.8 ± 2.3 | 77.7 ± 5.1 | 81.4 ± 3.5 | 96.1 ± 2.6 | 91.7 ± 2.6 |
| LEnALL | 89.9 ± 2 | 88.9 ± 2.4 | 98.7 ± 1.9 | 79 ± 4.8 | 82.6 ± 3.3 | 98.5 ± 2.3 | 91.6 ± 2.8 |
| LNLEnALL | 90.1 ± 2 | 89.2 ± 2.4 | 98.4 ± 1.9 | 79.9 ± 4.9 | 83.2 ± 3.3 | 98.2 ± 2.2 | 93.6 ± 2.3 |