一种基于神经网络的网络流量组合预测模型

被引:28
作者
冯海亮
陈涤
林青家
陈春晓
机构
[1] 山东大学信息科学与工程学院
关键词
流量; 相关; 非平稳; 神经网络; 预测;
D O I
暂无
中图分类号
TP393.06 [];
学科分类号
摘要
Internet流量是具有相关和非平稳特性的时间序列,文中通过对过去一些流量模型的分析,构建了一种新的预测模型。该模型首先使用小波方法对网络流量进行预处理,然后分别使用线性神经网络和E lman神经网络进行预测,以保证能够描述流量的相关和非平稳性,最后把两种预测的结果通过BP神经网络合成为最终预测结果。通过对不同流量的一步预测和多步预测仿真,验证了组合模型比单一模型具有更高的预测精度。
引用
收藏
页码:2206 / 2208+2228 +2228
页数:4
相关论文
共 8 条
[1]  
Multiresolution FIR neural-net-work-basedlearning algorithmapplied to network traffic predic-tion. V Alarcon-Aquino,J A Barria. IEEE Trans on Systems,Man,and Cybernetics—Part C:Applications and Reviews . 2006
[2]  
Study on Network Traffic Prediction Techniques. FENG HA,SHLT YT. Wireless Communications, Networking and Mobile Computing Proceedings . 2005
[3]  
Wavelet Networks. Qinghua Zhang,Albert Benveniste. IEEE Trans Nneural Network . 1992
[4]  
Dynam ic properties of E lman andmod ified E lman neural network. CHENG YC,QIWM,CAIWY. International Conference on Ma-ch ine Learn ing and Cybernetics . 2002
[5]  
Dynamic Bandwidth Allocation Using A Two-Stage Fuzzy Neural Network Based Traffic Predictor. SADEK N,KHOTANZAD A. International Joint Conference on Neural Networks, IJCNN’04 . 2004
[6]  
Traffic modeling, prediction and congestion control for high-speed networks: A Fuzzy AR Approch. CHEN BS,PENG SC,WANG KC. IEEE Transactions on Fuzzy Systems . 2000
[7]  
Mu lti-scale h igh-speed network trafficpred iction using comb ination of neural networks. KHOTANZAD A,SADEK N. Proceed ings ofthe International Joint Conference on Neural Networks . 2003
[8]  
Traffic Pred iction Using FAR IMAModels. SHU Y,JIN Z,ZHANG L,et al. ICC 99 . 1999