Fuzzy-ART in network anomaly detection with feature-reduction dataset

Author

Ngamwitthayanon, Nawa ; Wattanapongsakorn, Naruemon

Author_Institution

Dept. of Comput. Eng., Rajamangala Univ. of Technol. Isan, Khonkaen, Thailand

fYear

2011

fDate

26-28 Sept. 2011

Firstpage

116

Lastpage

121

Abstract

The validation of Fuzzy-Adaptive Resonance Theory (Fuzzy-ART or F-ART) was made in our work on Network Anomaly Intrusion Detection (NAID) application. Feature reduction of KDD 99 dataset was applied to the F-ART model and produced superior performance. We found the effectiveness of FART on clustering data instances into normal and anomalous traffic. The detection performance was clearly improved compare to the detection with the full-feature dataset. The results validated the capability of F-ART with one shot fast learning on the effectiveness of this adaptive learning algorithm along with the robustness and fast response that can provide a real-time network anomaly detection.

Keywords

adaptive resonance theory; computer network security; fuzzy set theory; learning (artificial intelligence); pattern clustering; KDD 99 dataset; adaptive learning algorithm; data clustering; feature-reduction dataset; fuzzy adaptive resonance theory; fuzzy-ART model; real time network anomaly intrusion detection application; Adaptation models; Adaptive systems; Clustering algorithms; Computational modeling; Data models; Feature extraction; Intrusion detection;

fLanguage

English

Publisher

ieee

Conference_Titel

Networked Computing (INC), 2011 The 7th International Conference on

Conference_Location

Gyeongsangbuk-do

Print_ISBN

978-1-4577-1129-9

Electronic_ISBN

978-89-88678-43-5

Type

conf

Filename

6058956