Distributed Intrusion Detection System Based on BP Neural Network

The central processing units of centralized structure are generally overloaded, and traditional intrusion detection system cannot effectively detect unknown attacks. To overcome the above problems, a distributed intrusion detection system model is established combining neural network with distributed detection in this paper based on the self-learning and adaptive characteristics of neural networks. A simulation experiment is done with Cauchy error estimation for avoiding trapping into local minimum. The result shows that the system can detect most of known attacks and analyze the unknown attacks, which is beneficial to artificial analysis and detection.