A new intrusion detection method
Ang Li · Journal of Nanyang Normal University · 2007
A method which a new clustering arithmetic RBF neural network is used in the intrusion detection is presented.Two-phase learning is considered in the method.When the network hidden layer centers are confirmed by unsupervised learning arithmetic,a strategybased on Gaussian basis and input-output clustering is presented.Based on Fisher detachable ratio,the punishment gene in Gaussian basis distance is designed,the clustering performance is improved.An intrusion detection model is built according to the method which is used in the paper,one hand,network training speed is enhanced,on the other hand,the intrusion detection performance in misinformation forecast is improved.The simulations result shows that the network has better detection effect.