Table 7 Comparative performance metrics of predictive models in public data and validation cohort
Performance metric | Public data (n = 1,000) | Validation cohort (n = 10) |
|---|---|---|
AUC (95% CI) | 0.89 (0.87–0.91) | 0.82 (0.72–0.92) |
Sensitivity | 88% | 85% |
Specificity | 90% | 83% |
Accuracy | 89% | 84% |
Positive predictive value | 89% | 83% |
Negative predictive value | 89% | 85% |
F1 score | 0.885 | 0.840 |
Brier score | 0.072 | 0.135 |