Title :
An improved intrusion detection system based on neural network
Author_Institution :
Sch. of Software & Microelectron., Peking Univ., Beijing, China
Abstract :
Neural network approach is an advanced methodology used for intrusion detection. Adaptive Resonance Theory (ART) is one kind of neural networks featuring self-organization of stable recognition categories and on-line learning. ART 2-A is a fast clustering algorithm of the ART family. But ART 2-A cannot process categorical values, while network traffic data collected for intrusion detection always contain this kind of data. An improved model of ART 2-A is thus proposed to deal with categorical data properly. To verify the accuracy of this new model, experiments were carried out on the KDD Cup 99 data set. Results showed that the performance of the new model was satisfactory.
Keywords :
ART neural nets; pattern clustering; security of data; ART 2-A clustering algorithm; KDD Cup 99 data set; adaptive resonance theory; intrusion detection system; neural network; online learning; stable recognition category; Clustering algorithms; Feedforward neural networks; Intrusion detection; Microelectronics; Neural networks; Protection; Resonance; Subspace constraints; Telecommunication traffic; Traffic control; information security; intrusion detection; neural network;
Conference_Titel :
Intelligent Computing and Intelligent Systems, 2009. ICIS 2009. IEEE International Conference on
Conference_Location :
Shanghai
Print_ISBN :
978-1-4244-4754-1
Electronic_ISBN :
978-1-4244-4738-1
DOI :
10.1109/ICICISYS.2009.5358048
