融合EMD与全局版人工鱼群LS-SVM的网络流量预测

发布时间：2018-03-31 15:18

本文选题：经验模式分解　切入点：本征模式分量　出处：《激光杂志》2014年05期

【摘要】：为提高网络流量预测模型对稳定性和精度要求,提出了一种融合经验模式分解(EMD)和全局版人工鱼群LS-SVM模型的网络流量预测算法。该模型首先通过EMD分解,将原始流量的非平稳序列分解为若干不同尺度的平稳IMF(固有模态分量)分量,再用LS-SVM模型分别对各分量进行预测,其中LS-SVM的参数用全局版人工鱼群算法进行优化;最后通过SVM组合得到原始序列的预测值。仿真实验验证了该算法的预测准确性和稳定性。
[Abstract]:In order to improve the stability and accuracy of network traffic prediction model, a network traffic prediction algorithm based on empirical mode decomposition (EMD) and global artificial fish swarm LS-SVM model is proposed. The model is decomposed by EMD. The non-stationary sequence of the original flow is decomposed into a number of stationary IMF (intrinsic mode component) components of different scales, and then each component is predicted by LS-SVM model, in which the parameters of LS-SVM are optimized by the global version artificial fish swarm algorithm. Finally, the prediction value of the original sequence is obtained by SVM combination, and the simulation results show that the algorithm is accurate and stable.
【作者单位】：新疆大学信息科学与工程学院;中国移动通信集团新疆有限公司;
【基金】：中国移动通信集团新疆有限公司发展基金项目(项目编号:XJM2012-01)
【分类号】：TP393.06

【参考文献】