Network Intrusion Detection based on RBF Neural Networks and Fuzzy Cluster

Network Intrusion detection is a key research topic in the field of information security. In view of the shortcomings of high data dimension and low detection accuracy for traditional detection algorithm, a detection algorithm is proposed which combined fuzzy clustering and RBF neural network. The original data set is reduced effectively by fuzzy clustering algorithm, while optimal model of RBF neural network is selected by taking of the method of cross-validation. Experiments includes the intrusion data reduction, classifier optimization, algorithm accuracy and its time consumption. The results show that the proposed algorithm in this paper can effectively reduce the original data set and its classification accuracy rate of more than 90%, since the overall algorithm performs well.