朱润涛， 李宁^*

（  北京邮电大学电子工程学院；  ）

摘要： 针对卫星网络星上处理能力有限的特点，提出一种基于小波变换和非线性优化SVM的网络流量预测模型。首先对网络流量进行小波变换得到尺度系数和小波系数。再利用非线性量子粒子群算法对SVM进行优化。通过SVM回归模型对小波分量进行回归分析，最后将小波进行重构得到预测结果。实验结果表明，改进过的预测算法具有较好的泛化能力，适合短时预测，预测精度明显好于传统预测方法。
关键词： 支持向量机；卫星网络；流量预测

Zhu Runtao， Li Ning^*

（  the School of Electronic Engineering, Beijing University of Posts and Telecommunications；  ）

Abstract： According to the limited processing ability of satellite, a network traffic prediction model based on wavelet transformation and nonlinear optimization SVM is proposed. First wavelet transform was carried out on the network traffic to get scale coefficients and the wavelet coefficients. Then Using nonlinear quantum particle swarm optimization algorithm to optimize the SVM. Through the SVM regression model, the wavelet component regression analysis is done. Finally do wavelet reconstruction and obtain prediction results. The experimental results show that the improved prediction algorithm has better generalization ability and is suitable for short-term prediction. Its prediction accuracy is much better than traditional forecasting methods.

Tag：

点此返回栏目查看更多>>>参考论文