摘要
为了提高流量预测的准确性,将混沌理论和在线LS-SVM回归技术应用于网络流量预测。采用相空间重构理论计算流量的时延τ、嵌入维数m,据此确定训练样本对并建立在线预测模型,对网络流量数据进行预测。结果表明,该方法能有效地进行流量预测,相对于传统的离线预测方式,该方法具有更好的预测精度。
The chaos theory and online LS-SVM regression are applied to network traffic flow prediction to improve its accuracy. The phase space reconstruction theory is introduced to calculate the delay time (-) and embedded dimension (m) . On this basis, the training samples are formed and the online LS-SVM prediction model is constructed to predict the network traffic flow. The results show that the online LS-SVM predition model can effectively predict the network traffic flow and get higher accuracy compared with the off-line prediction model.
出处
《电视技术》
北大核心
2012年第7期67-70,共4页
Video Engineering
基金
陕西省自然科学基金项目(SJ08F14
2009JQ8008)
关键词
混沌理论
最小二乘支持向量机
网络流量
在线预测
迭代算法
chaos theory
least squares support vector machines
network traffic flow
online prediction
iterative algorithm