首页 | 本学科首页   官方微博 | 高级检索  
     

网络流量短期预测方法的研究与应用
引用本文:谭晓玲,许勇,张凌,梅成刚,刘兰. 网络流量短期预测方法的研究与应用[J]. 计算机工程与设计, 2006, 27(8): 1341-1342,1345
作者姓名:谭晓玲  许勇  张凌  梅成刚  刘兰
作者单位:1. 重庆三峡学院,电子工程系,重庆,404000
2. 华南理工大学,广东省计算机网络重点实验室,广东,广州,510640
3. 华南理工大学,广东省计算机网络重点实验室,广东,广州,510640;广东技术师范学院,电子系,广东,广州,510655
摘    要:
给出了网络流量短期预测方法.该方法运用小波变换自适应时频局部化分析方法和改进的Mallat算法将网络流量分解到不同频带上,然后对各子频带上的小波分进行不同阈值的消噪处理,再对仍是非平稳过程的分量进行差分处理使其转化为平稳序列,最后对各平稳过程分量采用ARMA模型进行预测.实际流量分析表明该方法简便,且短期预测精度较高.

关 键 词:网络流量  小波分析  Mallat算法  ARMA模型  预测
文章编号:1000-7024(2006)08-1341-02
收稿时间:2005-01-31
修稿时间:2005-01-31

Research and application about method of network traffic short-term forecasting
TAN Xiao-ling,XU Yong,ZHANG Ling,MEI Cheng-gang,LIU Lan. Research and application about method of network traffic short-term forecasting[J]. Computer Engineering and Design, 2006, 27(8): 1341-1342,1345
Authors:TAN Xiao-ling  XU Yong  ZHANG Ling  MEI Cheng-gang  LIU Lan
Affiliation:1. Department of Electronic Engineering, Chongqing Three Gorges University, Chongqing 404000, China; 2. Guangdong Key Laboratory of Computer Network, South China University of Technology, Guangzhou 510640, China; 3.Department of Electronic Information Engineering,Guangdong Polytechnic Normal University, Guangzhou 510655,China
Abstract:
Method of network traffic short-term forecasting is presented. The network traffic is decomposed into different frequency bands by using method of analyzing the self-adaptive time-frequency localization of wavelet transform, and the improved Mallat algorithm. And then, via different thresholds, wavelet weights in different frequency bands are denoised. After that, wavelet weights still in unstable sequences are transformed into stable sequences by carrying on difference disposal. Finally, the ARMA model is taken to predict the weights of all stable sequences. Practice of network traffic analysis shows that the method is simple, applicable and has high accuracy for short-term prediction.
Keywords:network traffic   wavelet analysis   Mallat algorithm   ARMA model   forecast
本文献已被 CNKI 维普 万方数据 等数据库收录!
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号