Table 2.
Modeling results sorted by performance.
| Model | Area under the receiver operator characteristic curve (SE) |
| KNeighborsUnif-AutoGluon | 0.648 (0.011) |
| KNeighborsDist-AutoGluon | 0.649 (0.011) |
| NeuralNetFastAI-AutoGluon | 0.858 (0.013) |
| LightGBMLarge-AutoGluon | 0.867 (0.014) |
| NeuralNetMXNet-AutoGluon | 0.907 (0.008) |
| RandomForestGini-AutoGluon | 0.911 (0.007) |
| ExtraTreesGini-AutoGluon | 0.911 (0.009) |
| RandomForestEntr-AutoGluon | 0.917 (0.008) |
| ExtraTreesEntr-AutoGluon | 0.918 (0.007) |
| LightGBM-AutoGluon | 0.929 (0.007) |
| XGBoost-AutoGluon | 0.931 (0.006) |
| LightGBMXT-AutoGluon | 0.931 (0.005) |
| GRU-Mean | 0.932 (0.005) |
| GRU-Forward | 0.933 (0.006) |
| CatBoost-AutoGluon | 0.933 (0.005) |
| AutoGluon | 0.934 (0.005) |
| GRU-Simple | 0.935 (0.004) |
| GRU-D | 0.938 (0.004) |