A Research on Intrusion Detection Based on Support Vector Machines

Author

Fang, Xiaozhao ; Zhang, Wei ; Teng, Shaohua ; Han, Na

Author_Institution

Fac. of Comput., Guangdong Univ. of Technol., Guangzhou, China

fYear

2010

fDate

13-14 Oct. 2010

Firstpage

109

Lastpage

112

Abstract

Mass of the training samples and setting parameters of SVM artificially will affect badly the efficiency to find an optimal decision hyper plane for SVM. In this paper, FCM clustering algorithm and heuristic PSO algorithm are applied to Intrusion Detection. FCM clustering algorithm is designed to help SVM to find the optimal training samples from vast amounts of data; heuristic PSO algorithm is designed to find optimal parameters for SVM intelligently. The result of simulations run on the data of KDDCUP1999 shows that this approach can not only reduce the number of training samples and training time for SVM, but also detect unknown and known intrusions efficiently in the network.

Keywords

fuzzy set theory; particle swarm optimisation; security of data; support vector machines; clustering algorithm; fuzzy c-means clustering; heuristic PSO algorithm; intrusion detection; optimal decision hyperplane; particle swarm optimization; support vector machines; Algorithm design and analysis; Classification algorithms; Clustering algorithms; Intrusion detection; Support vector machines; Training; Training data; Fuzzy C-means Clustering; Particle Swarm Optimization; Support Vector Machines; Support vector;

fLanguage

English

Publisher

ieee

Conference_Titel

Communications and Intelligence Information Security (ICCIIS), 2010 International Conference on

Conference_Location

Nanning

Print_ISBN

978-1-4244-8649-6

Electronic_ISBN

978-0-7695-4260-7

Type

conf

DOI

10.1109/ICCIIS.2010.42

Filename

5629206