Detecting Cyber-Attacks on Wireless Mobile Networks Using Multicriterion Fuzzy Classifier with Genetic Attribute Selection

Table 4

The KDD’99 per-class performance of the proposed method with and without attribute selection (approximated to three decimal digits).

Normal/attack	Count	With attribute selection				Without attribute selection
		Precision	Recall		AUC	Precision	Recall		AUC
Normal	97278	0.999	0.999	0.999	1	0.990	0.989	0.989	0.990
Back	2203	1	0.999	0.999	0.999	0.987	0.987	0.987	0.989
Buffer_overflow	30	1	0.692	0.818	0.846	0.723	0.678	0.700	0.848
ftp_write	8	0.5	0.5	0.5	0.75	0	0.000	0.000	0.573
Guess_passwd	53	0.909	0.952	0.93	0.976	0.933	0.933	0.933	0.962
imap	12	1	0.4	0.571	0.7	0.157	0.240	0.190	0.670
ipsweep	1247	0.892	0.988	0.938	0.994	0.985	0.983	0.984	0.989
Land	21	0.8	1	0.889	1	0.847	0.937	0.890	0.919
Loadmodule	9	0.333	0.2	0.25	0.6	0	0.000	0.000	0.766
Multihop	7	0.25	0.333	0.286	0.667	0.276	0.323	0.298	0.847
Neptune	107201	1	1	1	1	0.990	0.990	0.990	0.990
nmap	231	0.906	0.358	0.513	0.679	0.938	0.981	0.959	0.981
Perl	3	0.5	0.5	0.5	0.75	0.323	0.990	0.490	0.980
phf	4	0.25	1	0.4	1	0.990	0.657	0.790	0.990
pod	264	0.986	0.973	0.98	0.987	0.990	0.986	0.988	0.990
Portsweep	1040	0.981	0.976	0.979	0.988	0.977	0.982	0.979	0.987
Rootkit	10	0.5	0.333	0.4	0.667	0	0.000	0.000	0.662
Satan	1589	0.986	0.989	0.988	0.995	0.981	0.984	0.982	0.987
Smurf	280790	1	1	1	1	0.990	0.990	0.990	0.990
Spy	2	1	1	1	1	0	0.000	0.000	0.443
Teardrop	979	0.994	0.997	0.996	0.999	0.989	0.990	0.989	0.989
Warezclient	1020	0.997	0.982	0.99	0.991	0.968	0.982	0.975	0.988
Warezmaster	20	0.333	0.5	0.4	0.75	0.790	0.752	0.770	0.910