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

Table 2

The comparative study of various ensemble methods. Different performance measures are used to study their performance.

Ensemble	Accuracy	Precision	Recall	F1-measure	AUROC	AUPRC
Single decision tree	0.845	0.386	0.275	0.321	0.609	0.247
Random forest	0.878	0.585	0.300	0.397	0.872	0.499
Bagging	0.876	0.568	0.313	0.403	0.858	0.506
XGBoost	0.865	0.533	0.575	0.536	0.863	0.505
AdaBoost	0.860	0.462	0.300	0.364	0.844	0.410
Ensembles for imbalanced datasets
Balanced random forest (RUS)	0.816	0.410	0.812	0.540	0.879	0.561
SmoteBagging	0.858	0.518	0.512	0.493	0.845	0.486
RUSBagging	0.835	0.432	0.675	0.517	0.881	0.516
SmoteBoost	0.868	0.544	0.562	0.530	0.849	0.514
RUSBoost	0.781	0.316	0.562	0.397	0.716	0.340

Bold numbers show the best performances.