Table 2 Multi-model classification—validation set results.
Model | AUC (SD) | Cut-off (SD) | Accuracy (SD) | Sensitivity (SD) | Specificity (SD) | Positive predictive value (SD) | Negative predictive value (SD) | F1 score (SD) | Kappa (SD) |
|---|---|---|---|---|---|---|---|---|---|
XGBoost | 0.808 (0.022) | 0.863 (0.011) | 0.817 (0.020) | 0.680 (0.045) | 0.865 (0.018) | 0.836 (0.064) | 0.814 (0.015) | 0.748 (0.044) | 0.437 (0.071) |
Logistic | 0.747 (0.041) | 0.328 (0.031) | 0.767 (0.027) | 0.609 (0.075) | 0.832 (0.043) | 0.574 (0.055) | 0.844 (0.018) | 0.586 (0.036) | 0.410 (0.048) |
LightGBM | 0.818 (0.022) | 0.876 (0.009) | 0.807 (0.022) | 0.709 (0.070) | 0.842 (0.017) | 0.842 (0.071) | 0.803 (0.017) | 0.769 (0.064) | 0.394 (0.081) |
RandomForest | 0.797 (0.022) | 0.450 (0.032) | 0.805 (0.018) | 0.683 (0.057) | 0.827 (0.030) | 0.680 (0.050) | 0.838 (0.009) | 0.681 (0.052) | 0.463 (0.040) |
SVM | 0.732 (0.047) | 0.273 (0.014) | 0.713 (0.036) | 0.620 (0.077) | 0.792 (0.049) | 0.475 (0.050) | 0.833 (0.021) | 0.537 (0.060) | 0.321 (0.071) |
KNN | 0.690 (0.030) | 0.400 (0.000) | 0.776 (0.013) | 0.474 (0.135) | 0.840 (0.081) | 0.662 (0.046) | 0.795 (0.010) | 0.542 (0.103) | 0.324 (0.044) |
