Feature Entropy Estimation (FEE) for Malicious IoT Traffic and Detection Using Machine Learning

Table 2

Evaluation of proposed methodology on NSL-KDD dataset.

Testsets		Accuracy	Precision	Recall	F_measure
Testset_1	DoS	94.54444	91.10397	94.22564	92.63851
	Probe	95.93165	84.76717	68.65544	75.86532
	R2L	99.50386	95.67901	38.94472	55.35714
	U2R	99.96626	99.97023	99.99603	99.98313
Testset_2	DoS	94.57939	91.14358	94.29584	92.69292
	Probe	95.12826	90.0986	53.4878	67.1258
	R2L	99.54866	94.96855	44.15205	60.27944
	U2R	99.96825	99.97505	99.99319	99.98412
Testset_3	DoS	94.64966	91.31289	94.31761	92.79093
	Probe	95.21327	90.21739	54.37197	67.85143
	R2L	99.51577	96.18321	41.44737	57.93103
	U2R	99.97354	99.97883	99.99471	99.98677
Testset_4	DoS	94.62738	91.27438	94.31897	92.7717
	Probe	94.47814	90.65934	45.15908	60.28774
	R2L	99.50783	94.17476	39.43089	55.58739
	U2R	99.98095	99.98412	99.99682	99.99047
Testset_5	DoS	94.57432	91.14699	94.23958	92.66749
	Probe	95.92379	84.40888	69.27966	76.0996
	R2L	99.51578	96.20253	38.97436	55.47445
	U2R	99.98015	99.98412	99.99603	99.99008