摘要
互联网流量数据属于非平稳的时间序列,具有很强的突发性和自相似性等分形特征。小波分析能够保持对象的尺度不变性,很适合分析和处理自相似过程。分析了常见流量模型的优缺点,利用小波技术将网络流量分解、重构,并结合传统FARIMA模型分析和预测网络流量,实验结果表明该方法可以有效地对网络流量进行建模和预测。
The Internet traffic belongs to non-stationary time series, and it has some fractal characteristics of strong bursty and self-similarity. The wavelet technology can preserve the scale invariable, and it is applied to analyze and deal with the self-similarity process. The advantage and disadvantage of the existing models are analyzed, and the network traffic is analyzed and forecasted the wavelet technology and the traditional FARIMA model. The experimental result shows that this method is effective for network traffic modeling and forecasting.
出处
《计算机应用与软件》
CSCD
北大核心
2008年第8期70-72,共3页
Computer Applications and Software
基金
陕西省自然科学基金项目(2005f36)
