Table 2.

Comparison of models for the confidential dataset

External validation using the confidential dataset (800 structures)
	BAC	SEN	SPEC	PPV	NPV
SOHN	0.67	0.53	0.80	0.55	0.80
Random Forest	0.64	0.50	0.77	0.50	0.77
SVM	0.64	0.50	0.77	0.50	0.77
kNN	0.63	0.48	0.79	0.51	0.79
Decision Tree	0.62	0.45	0.80	0.50	0.80

BAC = balanced accuracy, SEN = sensitivity, SPEC = specificity, PPV = positive predictivity, NPV = negative predictivity.