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

Table 7

The effect of age variable on the performance (AUROC) 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		0.867	0.872 (+)	0.878	0.882 (+)
Bagging		0.861 (+)	0.858	0.861 (+)	0.856
XGBoost		0.860	0.863 (+)	0.866	0.867 (+)
AdaBoost		0.850 (+)	0.844	0.865 (+)	0.850
Ensembles for imbalanced datasets
Balanced random Forest (RUS)		0.892 (+)	0.879	0.895 (+)	0.889
SmoteBagging		0.862 (+)	0.845	0.859 (+)	0.854
RUSBagging		0.883 (+)	0.881	0.887 (+)	0.880
SmoteBoost		0.863 (+)	0.849	0.857	0.863 (+)
RUSBoost		0.755 (+)	0.716	0.727 (+)	0.7171