基于WA-ELM的网络流量混沌预测模型

发布时间：2018-03-21 16:45

本文选题：网络流量　切入点：极限学习机　出处：《微电子学与计算机》2017年06期 　论文类型：期刊论文

【摘要】：针对当前网络流量预测模型存在的缺陷,结合网络流量的混沌特性,提出了小波变换和极限学习机的网络流量预测模型(WA-ELM).首先采用小波变换对网络流量时间序列进行处理,得到不同频率特征的分量,并对各特征分量进行相空间重构,然后采用极限学习机对网络流量各分量进行建模与预测,并对网络流量分量的预测值进行小波重构,得到原始网络流量的预测值,最后采用具体网络流量预测结果进行了验证,并与其他模型进行了对照测试.结果表明,WA-ELM获得了比其他模型更高的网络流量预测精度,而且网络流量的预测结果更加稳定,为网络流量提供了一种新的建模工具.
[Abstract]:In view of the shortcomings of the current network traffic prediction model and the chaotic characteristics of the network traffic, a network traffic prediction model of wavelet transform and extreme learning machine is proposed. Firstly, wavelet transform is used to deal with the network traffic time series. The components of different frequency features are obtained, and each characteristic component is reconstructed in phase space. Then, the network traffic components are modeled and predicted by extreme learning machine, and the predicted value of network traffic component is reconstructed by wavelet transform. The prediction value of the original network traffic is obtained, and the results of network traffic prediction are verified by the concrete network traffic prediction results, and compared with other models, the results show that WA-ELM has higher network traffic prediction accuracy than other models. Moreover, the prediction results of network traffic are more stable, which provides a new modeling tool for network traffic.
【作者单位】：河套学院理学系;
【分类号】：TP393.06

【相似文献】