Table 5. Predictive performance based on random forest for the training, test, and external dataset.
Dataset | Year | Number | Accuracy (%) | Sensitivity (%) | Specificity (%) | AUC |
---|---|---|---|---|---|---|
Training | 2007–2018 | 8644 | 83.39 | 83.02 | 83.76 | 0.9197 |
Test | 8644 | 64.78 | 66.58 | 64.16 | 0.7208 | |
External | 2019–2020 | 3352 | 62.98 | 68.55 | 60.97 | 0.7123 |