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

Table 3

The performance of the Standard decision tree ensembles with the SMOTE oversampling method for different ratios of minority and majority class points. AUROC and AUPRC performance measures are used in the experiments. Bold numbers show the best results.

Ensemble	Original		Ratio = 1		Ratio = 0.75		Ratio = 0.50
	AUROC	AUPRC	AUROC	AUPRC	AUROC	AUPRC	AUROC	AUPRC
Single decision tree	0.609	0.247	0.679	0.245	0.673	0.273	0.664	0.275
Random forest	0.872	0.499	0.863	0.480	0.869	0.502	0.86	0.492
Bagging	0.858	0.506	0.838	0.481	0.855	0.497	0.846	0.471
XGBoost	0.863	0.505	0.859	0.499	0.859	0.504	0.854	0.500
AdaBoost	0.844	0.410	0.823	0.431	0.847	0.482	0.834	0.434