Table 2.

Evaluation of machine learning algorithms: scalar performance measures and confusion matrix terms. Values are color-coded on a green (favorable values)-to-red (adverse values) scale.

	Accuracy	Precision	Recall	F1-Score	AUC	Time [s]
Model		PPV	TPR				TPR	TNR	FPR	FNR
Random Forest	0.960	0.946	0.916	0.930	0.947	0.437	0.92	0.98	0.02	0.08
Multi-layer Perceptron	0.944	0.899	0.911	0.905	0.934	24,130	0.91	0.96	0.04	0.09
k-Nearest Neighbors	0.921	0.890	0.832	0.860	0.895	0.001	0.83	0.96	0.04	0.17
SVM (linear kernel)	0.873	0.785	0.779	0.782	0.845	7.825	0.78	0.91	0.09	0.22
Naïve Bayes	0.851	0.752	0.734	0.743	0.817	0.004	0.73	0.90	0.10	0.27
Logistic Regression	0.842	0.816	0.595	0.688	0.770	0.042	0.59	0.94	0.06	0.41

Based on 80:20 split and fixed seed. PPV, positive predictive value; TPR, true positive rate; TNR, true negative rate; FPR, false positive rate; FNR, false negative rate.