Table 2.

Performance of the models in predicting the occurrence of MCE on the train and test sets.

Evaluation metric	Penalized Logistic Regression	Classification and regression trees	Random Forest	eXtreme Gradient Boosting
Train set
AU-ROC*	0.730	0.690	0.709	0.728
Test set
AU-ROC*	0.731	0.692	0.707	0.730
AU-PRC**	0.759	0.687	0.711	0.760
Accuracy	0.673	0.670	0.666	0.668
Specificity	0.563	0.562	0.527	0.522
Sensitivity	0.762	0.758	0.778	0.785
Precision	0.684	0.682	0.671	0.671
F1 score	0.721	0.718	0.721	0.723

*AU-ROC, Area Under the Receiver Operating Characteristic curve.

**AU-PRC, Area Under the Precision-Recall curve.