Table 2 ROC analysis of neural network output at different cutoff points in the testing set
Cutoff point | Sensitivity% | Specificity% | PPV% | NPV% | Youden Index% | Negatives correctly identified% |
|---|---|---|---|---|---|---|
0.015 | 100.0 | 89.3 | 13.2 | 100.0 | 89.3 | 87.8 |
0.02 | 95.7 | 92.3 | 16.9 | 99.9 | 88.0 | 90.9 |
0.04 | 93.6 | 96.2 | 28.4 | 99.9 | 89.8 | 94.7 |
0.06 | 91.5 | 97.4 | 36.4 | 99.9 | 88.9 | 96.0 |
0.08 | 89.4 | 98.1 | 42.9 | 99.8 | 87.5 | 96.6 |
0.1 | 87.2 | 98.6 | 50.6 | 99.8 | 85.8 | 97.2 |
0.3 | 70.2 | 99.8 | 86.8 | 99.5 | 70.0 | 98.7 |
0.5 | 63.8 | 99.9 | 90.9 | 99.4 | 63.7 | 98.9 |
0.6 | 63.8 | 100.0 | 100.0 | 99.4 | 63.8 | 99.0 |