网络流量行为分析的小波加权混沌局域模型

Wavelet weighted chaos local-region model of network traffic behavior analysis

结合小波变换和混沌局域模型的各自优势,提出一种网络流量的预测模型。首先,将网络流量时间序列进行小波分解得到高频信号序列和低频信号序列,再用加权混沌局域模型对每一成分的信号序列分别进行预测,对所有的预测分量进行小波重构就可以实现对网络流量的预测。用实际网络流量对该模型进行验证,实验结果表明,该模型具有较高的预测效果。

Integrating the advantage of wavelet transform with that of chaos local-region model, a new model of forecast network traffic was proposed. First, the network traffic time series was decomposed to the high frequency signal series and low frequency signal series and the weighted chaos local-region model was applied to predict these series respectively. Finally, forecasted traffic was achieved by wavelet reconstruction of all the forecasted components. The simulation results on real network traftlc indicate that this model is more satisfactory than traditional methods in network traffic prediction.