Application of a New Feature Generation Algorithm in Intrusion Detection System

Table 4

Multiclassification results of our method in the KDD99 dataset.

	Precision	Recall	F1-score	Support
smurf	1.0000	1.0000	1.0000	164091
imap	0.0000	0.0000	0.0000	1
land	1.0000	1.0000	1.0000	9
Normal.	1.0000	0.9990	0.9995	60593
ftp_write.	0.0000	0.0000	0.0000	3
rootkit.	0.7500	0.2308	0.3529	13
ipsweep.	1.0000	0.9804	0.9901	306
back	0.4015	1.0000	0.5729	1098
satan	0.9802	1.0000	0.9900	1633
neptune	1.0000	1.0000	1.0000	58001
teardrop	0.4444	1.0000	0.6154	12
multihop	1.0000	0.4444	0.6154	18
loadmodule	1.0000	1.0000	1.0000	2
portsweep	1.0000	1.0000	1.0000	354
warezmaster	1.0000	0.0031	0.0062	1602
nmap	1.0000	1.0000	1.0000	84
buffer_overflow	1.0000	1.0000	1.0000	22
perl	1.0000	1.0000	1.0000	2
pod	1.0000	1.0000	1.0000	87
phf	0.0000	0.0000	0.0000	2
guess_passwd	1.0000	1.0000	1.0000	4367
Accuracy			0.9942	292300
Macro avg.	0.7893	0.7456	0.7211	292300
Weighted avg.	0.9976	0.9942	0.9927	292300