[Retracted] A Double-Layered Belief Rule Base Model for Driving Anger Detection Using Human, Vehicle, and Environment Characteristics: A Naturalistic Experimental Study

Table 8

The average identification performances of the six models for four driving anger states.

	Dl_BRB	C4.5	BPNN	NBC	SVM	kNN
TPR	0.8433	0.8366	0.8277	0.8111	0.8015	0.7872
PPA	0.9187	0.9072	0.8938	0.8708	0.8651	0.8509
F1	0.8793	0.8704	0.8594	0.8398	0.8321	0.8178
FPR	0.0792	0.0689	0.0830	0.0894	0.0862	0.1033
Acc	0.8367	0.8252	0.8215	0.8014	0.7792	0.7625
Runtime ratio	0.8824	0.8715	1.0000	0.8965	0.9845	0.9136