MFVT: an anomaly traffic detection method merging feature fusion network and vision transformer architecture

EURASIP Journal on Wireless Communications and Networking

Table 3 Experimental results on the 2012 dataset

Methods	Precision	Recall	F1-socre	FPR	Accuracy
MFVT	0.9986	0.9975	0.998	0.000525	0.9988
MFVT (CPR)	0.9995	0.9994	0.9995	0.000175	0.9996
vision transformer	0.9984	0.9977	0.998	0.000625	0.9985
PCCN	0.9987	0.9979	0.9983	0.000575	0.9986
CNN	0.9958	0.9942	0.9949	0.00145	0.9962
CNN_LSTM	0.9949	0.9936	0.9942	0.001775	0.9951
DLSTM	0.9939	0.9928	0.9933	0.00195	0.9944
KNN	0.993	0.9903	0.9917	0.002125	0.9939
LR	0.9891	0.9902	0.9897	0.00315	0.9909
RF	0.9973	0.9966	0.9969	0.00085	0.9979
DT	0.9984	0.9984	0.9984	0.000375	0.999
SVM	0.9943	0.9937	0.994	0.0018	0.9949