Table 6.

Predictive performance of static baseline models for violent behavior.

Model type	Sensitivity	Specificity	Positive predictive value	Negative predictive value	Accuracy	AUC
MLP	0.4595	0.8806	0.6800	0.7468	0.7308	0.5484
RF	0.2973	0.8060	0.4583	0.6750	0.6250	0.6769
KNN	0.2973	0.8358	0.5000	0.6829	0.6442	0.6480
SVM	0.4865	0.8955	0.7200	0.7595	0.7500	0.7691
LR	0.4846	0.8955	0.7200	0.7595	0.7500	0.7953
XGBoost	0.4865	0.9104	0.7500	0.7625	0.7596	0.7749