Title Of Paper:
Intrusion Detection System with Hierarchical Different Parallel Classification
Author's Name :  Behrouz Safaiezadeh, Alireza Zebarjad, Amin Einipour
KeyWords:   Intrusion Detection System, support vector machine, neural network of multilayer Perceptron, fuzzy system
Pages:  39 -45
Volume: 3
Issue: 12
Year: 2015

Todays, lives integrated to networks and internet. The needed information is transmitted through networks. So, someone may attempt to abuse the information and attack and make changes by weakness of networks. Intrusion Detection System is a system capable to detect some attacks. The system detects attacks through classifier construction and considering IP in network. The recent researches showed that a fundamental classification cannot be effective lonely and due to its errors, but mixing some classifications provide better efficiency. So, the current study attempt to design three classes of support vector machine, the neural network of multilayer perceptron and parallel fuzzy system in which there are trained dataset and capability to detect two classes. Finally, decisions made by an intermediate network due to type of attack. In the present research, suggested system tested through dataset of KDD99 and results indicated appropriate efficiency 99.71% in average.  

Full Text: