Network Traffic Classification Based on SD Sampling and Hierarchical Ensemble Learning

Table 9

Classification performance of original datasets.

	Precision	Recall	F1 score	FNR	FPR
Original layer 1
Benign	0.9992	0.9980	0.9986	0.0008	0.0020
Abnormal	0.9980	0.9992	0.9986	0.0020	0.0008
Accuracy					0.9986
Macro avg	0.9986	0.9986	0.9986	0.0014	0.0014
Weighted avg	0.9986	0.9986	0.9986	0.0014	0.0014
Original layer2
Benign	0.9992	0.9980	0.9986	0.0008	0.0020
DoS hulk	0.9846	0.9988	0.9917	0.0154	0.0012
DDoS	1.0000	1.0000	1.0000	0.0000	0.0000
PortScan	0.9996	0.9996	0.9996	0.0004	0.0004
DoS goldeneye	1.0000	0.9988	0.9994	0.0000	0.0012
FTP-patator	0.9993	1.0000	0.9997	0.0007	0.0000
DoS slowloris	0.9948	0.9941	0.9944	0.0052	0.0059
DoS slowhttptest	0.9969	0.9969	0.9969	0.0031	0.0031
SSH-patator	1.0000	0.9988	0.9994	0.0000	0.0012
Bot	0.9979	0.9918	0.9949	0.0021	0.0082
Infiltration	1.0000	0.5556	0.7143	0.0000	0.4444
Heartbleed	1.0000	0.6667	0.8000	0.0000	0.3333
Web attack	0.9906	0.9813	0.9859	0.0094	0.0187
Accuracy					0.9978
Macro avg	0.9972	0.9369	0.9596	0.0028	0.0631
Weighted avg	0.9978	0.9978	0.9978	0.0022	0.0022