Network Traffic Classification Based on SD Sampling and Hierarchical Ensemble Learning

Table 13

The classification performance of each model in the three sampling modes.

	Accuracy	Precision	Recall	F1 score	FNR	FPR
KNN	0.9909	0.9390	0.8814	0.8953	0.0610	0.1186
DT	0.9971	0.9432	0.9367	0.9369	0.0568	0.0633
SVC	0.9394	0.8678	0.7709	0.7960	0.1322	0.2291
DNN	0.9359	0.8017	0.8523	0.7945	0.1983	0.1477
Random forest	0.9974	0.9953	0.9367	0.9585	0.0047	0.0633
XGBoost	0.9849	0.9906	0.9068	0.9333	0.0094	0.0932
XGBoost + RF	0.9978	0.9972	0.9369	0.9596	0.0028	0.0631
SMOTE + KNN	0.9906	0.9247	0.9004	0.8974	0.0753	0.0996
SMOTE + DT	0.9964	0.9776	0.9532	0.9615	0.0224	0.0468
SMOTE + SVC	0.9308	0.8759	0.8428	0.8360	0.1241	0.1572
SMOTE + DNN	0.9611	0.8583	0.8489	0.8422	0.1418	0.1511
SMOTE + random forest	0.9969	0.9494	0.9805	0.9606	0.0506	0.0195
SMOTE + XGBoost	0.9977	0.9952	0.9805	0.9868	0.0048	0.0195
SMOTE + (XGBoost + RF)	0.9979	0.9973	0.9892	0.9930	0.0027	0.0108
Random sampling + KNN	0.9895	0.9443	0.8980	0.9080	0.0557	0.1020
Random sampling + DT	0.9967	0.9954	0.9702	0.9802	0.0047	0.0298
Random sampling + SVC	0.9174	0.8651	0.8091	0.8049	0.1349	0.1909
Random sampling + DNN	0.9451	0.8502	0.8934	0.8497	0.1498	0.1066
Random sampling + random forest	0.9968	0.9806	0.9269	0.9473	0.0194	0.0731
Random sampling + XGBoost	0.9928	0.9901	0.9901	0.9900	0.0099	0.0099
Random sampling + (XGBoost + RF)	0.9977	0.9973	0.9800	0.9876	0.0027	0.0200
SD sampling + KNN	0.9895	0.9323	0.8979	0.9015	0.0677	0.1021
SD sampling + DT	0.9957	0.9571	0.9703	0.9632	0.0429	0.0297
SD sampling + SVC	0.9509	0.8696	0.8612	0.8320	0.1304	0.1388
SD sampling + DNN	0.9475	0.8559	0.9318	0.8750	0.1441	0.0682
SD sampling + random forest	0.9962	0.9903	0.9620	0.9711	0.0097	0.0380
SD sampling + XGBoost	0.9982	0.9960	0.9978	0.9969	0.0040	0.0022
SD sampling + (XGBoost + RF)	0.9975	0.9966	0.9977	0.9971	0.0034	0.0023