摘要：In order to overcome the separately selection advantages of traditional feature and RBF neural network parameter, increase accuracy rate of network’s intrusion detection, there came up with a research on neural network intrusion detection of improved particle swarm optimization. According to optimize the feature selection of network and RBF neural network parameter, established a neural network intrusion detection model of IPSO-RBF, made convergence and disturbance variation analysis on improving optimized particle swarm optimization, finally made a experimental simulation to this detection model, the result showed: compared with traditional models, the neural network intrusion detection of improved particle swarm optimization was faster and has a higher accuracy rate as well as efficiency.
