Table 2.
Performance evaluation of five algorithm models in predicting postoperative in-hospital mortality.
Testing cohort | |||||
---|---|---|---|---|---|
Metrics | XGBoost | RF | LR | SVM | NB |
AUC | 0.8504 | 0.8961 | 0.7893 | 0.7705 | 0.8024 |
F1 | 0.1214 | 0.1284 | 0.0955 | 0.1235 | 0.0904 |
Se | 0.8500 | 0.9500 | 0.7500 | 0.6500 | 0.8000 |
Sp | 0.8508 | 0.8422 | 0.8287 | 0.8910 | 0.8047 |
Pr | 0.0654 | 0.0688 | 0.0510 | 0.0682 | 0.0479 |
Acc | 0.8508 | 0.8435 | 0.8277 | 0.8881 | 0.8047 |
Acc, Accuracy; AUC, Area Under the Curve; F1, F1 score; LR, Logistic Regression; NB, Naive Bayes; Pr, Precision; RF, Random Forest; Se, Sensitivity; Sp, Specificity; SVM, Support Vector Machine; XGBoost, eXtreme Gradient Boosting.