Table 5.
The AUROCa, accuracy, sensitivity, and specificity of the 6 classification models.
| Model | AUROC | Accuracy | Sensitivity | Specificity |
| RFb | 0.91 | 0.84 | 0.63 | 0.92 |
| LightGBMc | 0.86 | 0.81 | 0.63 | 0.89 |
| XGBoostd | 0.83 | 0.78 | 0.55 | 0.89 |
| KNNe | 0.77 | 0.76 | 0.60 | 0.84 |
| SVMf | 0.88 | 0.81 | 0.47 | 0.95 |
| LRg | 0.81 | 0.77 | 0.52 | 0.85 |
aAUROC: area under the receiver operating characteristic curve.
bRF: random forest.
cLightGBM: light gradient boosting machine.
dXGBoost: extreme gradient boosting.
eKNN: k-nearest neighbor.
fSVM: support vertical machine.
gLR: logistic regression.