Decision Tree Ensembles to Predict Coronavirus Disease 2019 Infection: A Comparative Study

Table 8

The effect of age variable on the performance (AUPRC) of various ensembles. “+” sign shows the better performance of with age or without age.

Ensemble		Size = 50		Size = 100
		With age	Without age	With age	Without age
Random forest		0.541 (+)	0.499	0.575 (+)	0.523
Bagging		0.539 (+)	0.506	0.548 (+)	0.490
XGBoost		0.517 (+)	0.505	0.537 (+)	0.507
AdaBoost		0.485 (+)	0.410	0.527 (+)	0.441
Ensembles for imbalanced datasets
Balanced Random Forest (RUS)		0.606 (+)	0.561	0.602 (+)	0.585
SmoteBagging		0.552 (+)	0.486	0.546 (+)	0.473
RUSBagging		0.582 (+)	0.516	0.598 (+)	0.532
SmoteBoost		0.554 (+)	0.514	0.547 (+)	0.518
RUSBoost		0.333	0.340 (+)	0.329	0.361 (+)