摘要
对网络流量序列的准确预测的基础是进行流量判决门限阈值的合理调控,为了提高网络流量预测的精度,进行阈值调控优化设计。传统方法采用流量序列调频估计的方法进行阈值调控,算法不能准确估计流量序列的包络特征,预测精度不好。提出一种基于包络特征的网络流量预测阈值调控算法。通过不同类型的网络设备进行流量采集终端联网构建,采用自相关特征奇异分解对流量序列进行线性特征叠加运算,构建网络流量预测的时间序列模型。采用统计学数学方法对流量序列的延迟时间互信息进行阈值调控自适应建模,提取流量序列的包络特征。实现对网络流量预测的阈值调控,提高预测精度。仿真实验表明,该算法能实现对流量预测阈值的自适应调控,预测精度较传统方法提高37%,性能优越于传统算法。
The accurate prediction of network traffic based on the sequence of reasonable control flow decision threshold, in order to improve the precision of the network traffic prediction, threshold control optimization design. The traditional meth-ods for the estimation of the flow series FM threshold control, algorithm can not accurately estimate the envelope character-istic flow sequence, the prediction precision is not good. Put forward a kind of threshold adjusting algorithm of network traf-fic prediction based on the envelope characteristics. Flow collection terminal network construction through the network of different types of equipment, using autocorrelation characteristics of singular decomposition of linear feature overlay opera-tions on flow sequence, building a time series model of network traffic prediction. Using statistical mathematical methods of delay time of mutual information to flow sequence threshold control adaptive modeling, envelope extraction characteristic flow sequence is obtained. The realization of the threshold regulation of network traffic prediction, improve the precision of prediction. Simulation results show that the algorithm can realize the adaptive prediction control threshold to flow, the pre-diction accuracy than the traditional method to improve the performance of the algorithm is superior to the traditional 37%.
出处
《科技通报》
北大核心
2015年第10期67-69,共3页
Bulletin of Science and Technology
基金
重庆市教委科学技术研究项目(KJ1403209)
关键词
网络流量
预测
包络特征
network traffic
prediction
envelop
作者简介
邓荣（1977-）,女,重庆万州人,硕士,副教授,研究方向：计算机网络技术.
