Table 3 Comparison of model performance in terms of AUC, Brier’s score, calibration intercept and slope averaged over the last 7 to 30 days at the end of the observational period in the forecasting validation
Methods | Predictive Accuracy | Calibration | ||
|---|---|---|---|---|
AUC | Brier | Intercept | Slope | |
Including hospital use variables | ||||
logistic | 0.9819 | 0.0094 | 0.1031 | 1.2468 |
LASSO | 0.9842 | 0.0096 | -0.2558 | 1.1376 |
GAM | 0.9823 | 0.0095 | 0.4267 | 1.2265 |
LDA | 0.9849 | 0.0172 | -2.1364 | 0.5276 |
Tree | 0.9781 | 0.0098 | -0.1898 | 1.0332 |
RF | 0.9808 | 0.0096 | -0.2993 | 0.5798 |
XGBoost | 0.9866 | 0.0091 | 0.4799 | 1.2934 |
Excluding hospital use variables | ||||
logistic | 0.9472 | 0.0121 | 0.0712 | 1.0800 |
LASSO | 0.9453 | 0.0121 | -0.1516 | 0.9540 |
GAM | 0.9473 | 0.0121 | 0.1983 | 1.0987 |
LDA | 0.9331 | 0.0185 | -1.5685 | 0.5427 |
Tree | 0.9133 | 0.0124 | -0.2133 | 1.0171 |
RF | 0.9229 | 0.0130 | -1.0875 | 0.6017 |
XGBoost | 0.9487 | 0.0123 | 0.1079 | 1.0421 |