Method of evolutionary neural network-based intrusion detection

Intrusion detection is an important defence to protect the security of computer network systems. With an integrated technique of genetic algorithm and neural network, a method of evolutionary neural networks is proposed to perform intrusion detection in this paper. It is a robust enough, parallel and nonlinear dynamic processing system to satisfy requirements of real-time processing and prediction with high accuracy. With the new method, the structure of the neural network is optimized using a genetic algorithm. The obtained neural network model is thus used for intrusion detection and prealarm with high accuracy.