Table 2.
Performance comparison between the five models.
| Algorithms |
AUROC
(95% CI) |
Accuracy
(95% CI) |
Average Precision
(95% CI) |
Sensitivity
(95% CI) |
Specificity
(95% CI) |
PPV
(95% CI) |
NPV
(95% CI) |
|---|---|---|---|---|---|---|---|
| LR | 0.790 (0.705–0.875) |
0.837 (0.762–0.912) |
0.387 (0.192–0.582) |
0.334 (0.770–0.590) |
0.965 (0.906–1.000) |
0.743 (0.386–1.000) |
0.851 (0.798–0.904) |
| DT | 0.823 (0.779–0.867) | 0.874 (0.825–0.923) | 0.549 (0.394–0.704) | 0.718 (0.490–0.946) | 0.914 (0.874–0.954) | 0.682 (0.569–0.795) | 0.928 (0.874–0.983) |
| KNN | 0.624 (0.545–0.703) | 0.775 (0.725–0.825) | 0.233 (0.173–0.293) | 0.147 (0.380–0.283) | 0.935 (0.887–0.983) | 0.371 (0.115–0.627) | 0.811 (0.781–0.841) |
| GaussianNB | 0.832 (0.769–0.895) | 0.813 (0.770–0.856) | 0.300 (0.204–0.396) | 0.231 (0.108–0.354) | 0.961 (0.933–0.989) | 0.599 (0.350–0.848) | 0.831 (0.804–0.858) |
| XGBoost | 0.927 (0.086–0.968) | 0.918 (0.838–0.998) | 0.683 (0.400–0.966) | 0.729 (0.457–1.000) | 0.966 (0.908–1.000) | 0.855 (0.627–1.000) | 0.934 (0.869–0.999) |
LR, logistic regress; DT, decision tree; KNN, K nearest neighbor; GaussianNB, Gaussian naive bayes; XGBoost, extreme gradient boost; TPR, true positive rate; TNR, true negative rate; PPV: positive predicted value; NPV: negative predictive value.