The local-model of chaos network traffic based on mutual information and principal componential analysis

A novel adding-weight local-region model is developed for forecasting chaos network traffic. In this model, the neighbor phase points of forecast center phase point with their weight are determined based on mutual information and principal componential analysis in the reconstructed phase space. The result of simulation shows the presented model can greatly improve precision of network traffic forecasting when the embedded dimension of network traffic time series is high, compared with the traditional model, in which the neighbor phase points with their weight are determined by space distance.