Table 2.

Results of training a random forest classifier, a support vector machine classifier, and a multinomial Naïve-Bayes classifier using 10-fold cross-validation, compared with a traditional risk calculation method of multiple logistic regression.

Type	Accuracy	AUC (95% CI)	AUPRC	Sensitivity	Specificity	F1 score
Random forest classifier	0.75	0.71 (0.70–0.73)	0.33	0.44	0.82	0.61
Support vector machine classifier	0.73	0.71 (0.69–0.72)	0.34	0.52	0.78	0.62
Multinomial Naïve-Bayes classifier	0.64	0.66 (0.65–0.68)	0.23	0.60	0.64	0.56
Multiple logistic regression	0.83	0.70 (0.69–0.72)	N/A	0.36	0.87	0.36a

^aCalculated from previously published results (6).

AUC: Area under the curve.

AUPRC: Area under the precision-recall curve.