ARIMA模型在网络流量预测中的应用研究  

Application Research on Network Traffic Prediction Base on ARIMA

作　　者：张冉[1] 赵成龙[1] 

机构地区：[1]泰山职业技术学院,山东泰安271000

出　　处：《计算机仿真》2011年第2期171-174,共4页Computer Simulation

摘　　要：针对网络运行安全和可靠的要求,研究网络流量预测问题。网络流量具有高度自相似、时变性和非线性等时间序列特征,传统预测方法无法捕捉其时变性和自相似规律,导致预测精度比较低。为了提高网络流量的预测精度,在分析网络流量特征的基础上,提出一种基于ARIMA模型的网络流量预测方法。先采用差分法对网络流量原始数据平稳化处理,提取网络流量数据的自相似特征,然后将平稳后的数据利用能很好反映时变性和非线性的ARIMA模型对进行拟合和检验,建立网络流量的最优预测模型,最后根据获得最优预测模型对网络流量实例数据进行仿真预测。仿真结果表明,ARIMA模型的网络流量预测精度比其它预测模型要高,能够很好的反映网络流量的规律,在网络流量预测中有广泛应用前景。Network traffic prediction is the core problem in network congestion control and network traffic management of Network.Because the network traffic dates have the features of highly self-similar,time-varying and nonlinear,the traditional time series forecasting method cannot reflect the network traffic features and the prediction accuracy is very low.In order to improve the prediction precision of network traffic,a network traffic prediction method is proposed based on the ARIMA model in this paper.Firstly,the network is smoothed by difference method and can capture the self-similarity feature of original data,and then the time series features are captured by BARIMA,lastly the optimal network traffic prediction model is built.Simulation results show that the ARIMA model of network traffic prediction accuracy is higher than the traditional prediction precision and ARIMA has wide application prospects in network traffic prediction.

关 键 词：网络流量 时间序列 预测 自回归模型 

分 类 号：TP309[自动化与计算机技术—计算机系统结构]