Table 9.
Metrics for the training and testing (external) cohorts for the XGBoosta model.
Training data size | AUCb score | Sensitivity | Specificity | Precision | F1-score |
10,000 | 0.76 | 0.21 | 0.97 | 0.69 | 0.33 |
100,000 | 0.81 | 0.27 | 0.98 | 0.73 | 0.39 |
1,000,000 | 0.82 | 0.33 | 0.97 | 0.69 | 0.45 |
aXGBoost: extreme gradient boosting.
bAUC: area under the receiver operating characteristic curve.