基于非线性算法的网络业务流量预测

基于递归最小二乘支持向量机,提出了一种网络业务流量非线性预测算法。通过最小二乘支持量机首先将原始的网络流量数据映射到一个高维空间中,进而在这个高维空间中对流量数据进行预测,使得在低维空间中非线性预测转化为高维空间中的线性预测,提高了预测性能。仿真结果表明,预测误差能够维持在5%以内。

Nonlinear prediction algorithm of network traffic flow

Based on Least Squares Support Vector Machines,A nonlinear prediction method of network traffic flow was proposed.The network traffic flow data was mapped into a high dimension space utilizing Least Squares Support Vector Machines,and the prediction was carried out in this space.The prediction performance was increased since the nonlinear prediction in low dimension space was transformed into a linear prediction in a high dimension space.The simulation results show that the prediction error can keep within 5%.