Table 2 Performances of all the models
| Â | AUC | High sensitivity operating point | High specificity operating point | ||||
|---|---|---|---|---|---|---|---|
Sensitivity | Specificity | Accuracy | Sensitivity | Specificity | Accuracy | ||
WSDL | 0.799 (0.712,0.883) | 0.826 (0.669,0.980) | 0.635 (0.576,0.694) | 0.651 (0.596,0.706) | 0.783 (0.609,0.942) | 0.701 (0.642,0.756) | 0.708 (0.653,0.760) |
baseline-DL | 0.748 (0.638,0.856) | 0.826 (0.671,0.984) | 0.518 (0.457,0.579) | 0.544 (0.486,0.601) | 0.681 (0.496,0.869) | 0.701 (0.642,0.758) | 0.699 (0.644,0.754) |
LR | 0.737 (0.632,0.839) | 0.826 (0.670,0.980) | 0.557 (0.497,0.618) | 0.579 (0.523,0.636) | 0.609 (0.407,0.807) | 0.701 (0.646,0.758) | 0.693 (0.639,0.749) |
XGBoost | 0.703 (0.584,0.821) | 0.826 (0.678,0.981) | 0.380 (0.324,0.437) | 0.418 (0.364,0.471) | 0.638 (0.456,0.835) | 0.701 (0.644,0.757) | 0.695 (0.641,0.750) |
SVM | 0.626 (0.524,0.735) | 0.826 (0.670,0.980) | 0.302 (0.251,0.351) | 0.346 (0.297,0.393) | 0.391 (0.238,0.566) | 0.701 (0.643,0.755) | 0.675 (0.620,0.727) |
RF | 0.619 (0.516,0.751) | 0.826 (0.671,0.982) | 0.315 (0.259,0.370) | 0.358 (0.305,0.410) | 0.464 (0.27,0.659) | 0.729 (0.675,0.782) | 0.707 (0.655,0.758) |
KNN | 0.566 (0.496,0.637) | 0.638 (0.512,0.761) | 0.506 (0.472,0.541) | 0.517 (0.485,0.550) | 0.638 (0.512,0.761) | 0.506 (0.472,0.541) | 0.517 (0.485,0.550) |