基于灰色小波的网络流量组合预测模型  

Combination prediction model to network traffic based on grey-wavelet

作　　者：巩林明[1,2] 张振国[1] 

机构地区：[1]陕西科技大学,陕西西安710021 [2]石河子大学,新疆石河子832000

出　　处：《计算机工程与设计》2010年第8期1660-1662,1677,共4页Computer Engineering and Design

摘　　要：为了更准确的预测网络流量和刻画流量特性,提出了一种基于小波变换理论与灰色理论的组合预测方法。该方法拟合了小波分解的平稳化处理优点和灰色预测中的平移处理优点,根据提出的预测方法建立了基于灰色小波的网络流量预测模型。首先对分解后的时间序列分别进行平移预处理,然后对预处理后的序列数据进行预测并重构,从而得到原始时间序列的预测值。对实际的采集数据进行仿真实验,仿真结果表明,该方法预测效果良好,优于一般小波预测方法。For more accurate to forecast the network traffic and portray the characteristics of network traffic, one kind of combination prediction method based on the theory of wavelet transform and grey is proposed, the merit of smooth transaction to the wavelet and the virtue of parallel movement to the grey prediction is fitted in this method. Prediction model of grey-wavelet is established according to this prediction method, first of all, the time series decomposed is removed flatly by separately, and then forecast and recomposition is processed to the sequence data, thus the forecast value of primitive time series is obtained. And through carrying on the modeling and the forecast to the actual gathered data, the favorable effect of this method is turned up, and this method surpassed the method of common prediction used wavelet.

关 键 词：小波变换 灰色预测 组合预测 GM(l 1)模型 时间序列 

分 类 号：TP301[自动化与计算机技术—计算机系统结构] TP393.01[自动化与计算机技术—计算机科学与技术]