Research Article
Mortality Prediction Using SaO2/FiO2 Ratio Based on eICU Database Analysis
Table 2
Results from the cross-validation folds from the hyperparameter tuned training set.
| ā | Accuracy | AUC | Recall | Prec. |
| 0 | 0.8817 | 0.8507 | 0.1815 | 0.6395 | 1 | 0.8847 | 0.8493 | 0.1947 | 0.6782 | 2 | 0.8847 | 0.8459 | 0.2013 | 0.6703 | 3 | 0.8788 | 0.8325 | 0.1842 | 0.5957 | 4 | 0.8851 | 0.8334 | 0.2072 | 0.6774 | 5 | 0.8847 | 0.8472 | 0.2336 | 0.6455 | 6 | 0.8809 | 0.8398 | 0.2237 | 0.6018 | 7 | 0.8788 | 0.8445 | 0.1612 | 0.6125 | 8 | 0.8800 | 0.8305 | 0.1809 | 0.6180 | 9 | 0.8898 | 0.8532 | 0.2500 | 0.7037 | Mean | 0.8829 | 0.8427 | 0.2018 | 0.6443 | SD | 0.0033 | 0.0077 | 0.0259 | 0.0351 |
|
|
AUC: area under the curve; SD: standard deviation; Prec.: precision. Final scores are means of the ten folds.
|