Table 5. Comparison of the predictive performance of XGBoost reduced models on training, validation, and testing data.

Data	Sample (N)	Accuracy	Miss-classification	AUC	Sensitivity	Specificity	Precision
Training	579	1.0000	0.0000	1.0000	1.0000	1.0000	1.0000
Validation	96	0.9792	0.0208	0.9924	0.9821	0.9750	0.9821
Testing	288	0.9722	0.0278	0.9735	0.9630	0.9841	0.9873