基于小波包消噪和Elman的网络流量预测

Network Traffic prediction Based on Wavelet Packet de-noising and Elman neural network

网络流量预测对于大规模网络的规划设计和网络资源管理等方面都具有积极的意义,是网络流量工程重要组成部分。结合小波包消噪和Elman神经网络的优势,先将原始流量序列进行小波包消噪,将消噪后的序列作为Elman神经网络的输入,待预测序列作为输出。通过前N天的流量序列,预测出后M天的流量序列。这里采用序列的前N天的数据为滑动窗,并将其映射为在该窗之后的M天的预测值。仿真实验表明,与未进行小波包消噪而直接利用Elman神经网络的模型比,基于小波包消噪和Elman神经网络的网络流量预测模型具有更好的预测能力。

Network traffic prediction is positive significance for design of network and resource management,and it is an important section of traffic engineering.This paper presents a method of network traffic prediction based on wavelet packet de-noising and Elman neural network.Firstly,de-noising the traffic time series with wavelet packet,then taking the de-noised series as the input of Elman neural network while the predictive series as the output of Elman neural network.By using the N days ＇ de -noised traffic time series to forecast the later M days＇ predictive series.Taking the N days＇ de-noised series as a sliding window,mapped into the later M days＇ predictive series.The emulation experiment results show that compared with the model only based on Elman neural network,the model based on wavelet packet de-noising and Elman neural network is more successful in the network traffic prediction.