Table 3.
Model performance evaluation using training and validation cohorts.
| Cohort | Model | AUC (95%CI) | Accuracy | Sensitivity | Specificity |
|---|---|---|---|---|---|
| Training | LR | 0.837 (0.784–0.889) | 0.828 | 0.552 | 0.935 |
| KNN | 0.992 (0.985–0.998) | 0.904 | 0.658 | 1.000 | |
| SVM | 0.781 (0.728–0.834) | 0.934 | 0.765 | 1.000 | |
| DT | 0.827 (0.772–0.883) | 0.808 | 0.623 | 0.881 | |
| RF | 1.000 (1.000–1.000) | 1.000 | 1.000 | 1.000 | |
| XGB | 0.884 (0.844–0.925) | 0.828 | 0.917 | 0.600 | |
| ANN | 1.000 (1.000–1.000) | 0.759 | 0.548 | 0.905 | |
| Validation | LR | 0.824 (0.737–0.912) | 0.802 | 0.444 | 0.932 |
| KNN | 0.792 (0.683–0.901) | 0.802 | 0.407 | 0.946 | |
| SVM | 0.677 (0.578–0.775) | 0.772 | 0.259 | 0.959 | |
| DT | 0.675 (0.559–0.791) | 0.703 | 0.444 | 0.797 | |
| RF | 0.858 (0.783–0.932) | 0.802 | 0.518 | 0.905 | |
| XGB | 0.780 (0.686–0.874) | 0.693 | 0.77 | 0.481 | |
| ANN | 0.858 (0.782–0.932) | 0.594 | 0.365 | 0.837 |
LR, logistic regression; KNN, K-nearest neighbor; SVM, support vector machine; DT, decision tree model; RF, random forest; XGBoost, extreme gradient boosting; ANN, artificial neural network.