Table 1.
Model | ACC | Precision | Recall | F1 | NPV | AUC | AUPR |
---|---|---|---|---|---|---|---|
M1 | 0.766 | 0.727 | 0.669 | 0.697 | 0.788 | 0.842 | 0.770 |
M2 | 0.741 | 0.712 | 0.603 | 0.652 | 0.758 | 0.820 | 0.740 |
M3 | 0.702 | 0.662 | 0.532 | 0.590 | 0.721 | 0.766 | 0.670 |
M1 + M2 | 0.805 | 0.776 | 0.725 | 0.750 | 0.823 | 0.883 | 0.828 |
M1 + M3 | 0.779 | 0.733 | 0.711 | 0.722 | 0.809 | 0.861 | 0.800 |
M2 + M3 | 0.799 | 0.761 | 0.729 | 0.745 | 0.822 | 0.873 | 0.813 |
M1 + M2 + M3 | 0.807 | 0.778 | 0.728 | 0.752 | 0.824 | 0.886 | 0.834 |
M1, M2, and M3 represent the base models ECCNN, Roost, and Magpie respectively.