Razieh Baradaran - Department of information technology, university of Qom, Qom, Iran
Mahdieh Hajimohammad Hosseini - Department of information technology, university of Qom, Qom, Iran

In today’s world that information is a great wealth,protection of this wealth has a certain importance. On the other hand, with spread of global internet network and growing use of it, information security has become more vulnerable.In this paper, we present a support vector machine based intrusion detection system that use BN-KDD data set for train and test. This data set is provided from science and technology university data mining lab that in fact it is an improvement form of KDD CUP 99. Mentioned system classifies each instancewhich is expressed with 41 features as normal or attack and has over 90 percent accuracy in experiments on training and testing data.

Intrusion detection system, data mining, support vector machine, BN-KDD