Incremental Wavelet Neural Network based Prediction of Network Security Situation

Abstract

Network security situation prediction can help network administrators to make security decisions, and the current network security technology cannot predict the future network security situation. Hereby, a network security situation prediction based on incremental wavelet neural network (IWNN) was proposed, which establishes the nonlinear mapping according to the time relationship of network security situation value. Meanwhile, in this method, the complexity of the future network security situation is considered by incremental learning. Therefore, IWNN has better generalization performance in network security situation prediction when considering the changes in the pattern of the future network attack, the network scale and the network security technology, etc. Finally, the performance of the proposed method is verified by the experimental data processing results.