论文部分内容阅读
网络流量预测是网络性能管理的一个重要组成部分,一种好的预测模型能比较准确地判断网络流量的发展趋势,对网络管理起到推进作用。提出了将变尺度法应用于指数平滑模型中,以预测误差平方和(SSE)最小作为目标,构造并自动生成了最佳平滑参数,使网络流量的预测模型得以优化,增强了指数平滑模型对时间序列的适应能力,较好地解决了指数平滑预测模型中,平滑参数靠检验确定且为静态,平滑初值难以确定并导致预测偏差等问题。通过分析,证明了此模型能够较准确地预测出网络的流量,从而提高了网络的服务质量。
Network traffic prediction is an important part of network performance management. A good prediction model can accurately judge the trend of network traffic and play a promoting role in network management. In this paper, a new method is proposed to apply the variable-scale method to the exponential smoothing model. The optimal smoothing parameters are constructed and automatically generated with the goal of minimizing the square error of prediction (SSE). The prediction model of network traffic is optimized and the exponential smoothing model is enhanced Time series of adaptive ability to better solve the exponential smoothing prediction model, the smoothing parameters determined by the test and static, smooth initial value is difficult to determine and lead to bias and other issues. The analysis proves that the model can predict the network traffic more accurately and improve the service quality of the network.