Abstract:

Network traffic has characteristics of nonlinearity, heteroskedasticity, and volatility clustering in some network environments. Traditional models such as Auto Regressive Moving Average(ARMA) model with wavelet transform fail to describe these characteristics very well. Therefore, Generalized Auto Regressive Conditional Heteroskedasticity(GARCH) model with wavelet transform is studied for network traffic forecast. With wavelet transform, a network traffic series is divided into low frequency part and high frequency part, which are applied for GARCH modeling and forecasting respectively. Then, the forecast results of the subseries are reconstructed so as to implement the forecast of the original network traffic. The simulation shows that the accuracy of the forecast of GARCH model with wavelet transform is much better than that of ARMA model with wavelet transform.

Key words: wavelet transform;ARMA model;GARCH model;network traffic forecast