Table 15 Prediction model results stratified across sex and age groups for death (F1-w is the weighted F1 score).
Models | AUC | Accuracy | F1-w | Sensitivity |
|---|---|---|---|---|
Sex | ||||
Male | ||||
Logistic regression | 72.4 | 64.4 | 67.0 | 71.2 |
XGBoost | 74.0 | 64.2 | 66.8 | 74.6 |
Female | ||||
Logistic regression | 72.9 | 67.9 | 70.8 | 65.1 |
XGBoost | 74.0 | 66.3 | 69.6 | 70.4 |
Age | ||||
20–40 | ||||
Logistic regression | 67.8 | 93.4 | 90.2 | 0.0 |
XGBoost | 70.2 | 93.1 | 90.5 | 4.2 |
40–60 | ||||
Logistic regression | 62.9 | 77.8 | 76.2 | 18.1 |
XGBoost | 65.0 | 74.3 | 75.2 | 32.1 |
60–80 | ||||
Logistic regression | 61.7 | 47.6 | 46.4 | 85.8 |
XGBoost | 63.6 | 46.9 | 44.9 | 89.2 |
