Modeling and forecasting for network traffic based on wavelet decomposition
基于小波分解的网络流量时间序列建模与预测

This paper proposed a network traffic forecasting methods based on wavelet decomposition and time series analysis method. Firstly, the method decomposed the network traffic time series in multiple stationary components by wavelet decomposition, then used the autoregressive moving average method to model the each stationary component separately. Finally combined all the components of the model to get the forecasting model of the original non-stationary network traffic time series. It carried out the simulation experiment on time series data of the network library. The simulation results show that, the proposed method improves the network traffic time series forecasting accuracy rate, and it is an efficient, robust network traffic forecasting method.