In light of the problem of LSSVM parameters optimisation,in this paper we propose a network traffic prediction model which is based on optimising LSSVM by mutation particle swarm optimisation (MPSO-LSSVM).First,the phase space reconstruction is made on network traffic sequence to construct the learning samples of least square support vector;then the mutation particle swarm optimisation is used to select the parameters of LSSVM so as to build the optimal network traffic prediction model;finally,the contrast experiment is carried out between it and other models.Comparison results show that with respect to contrast models,the MPSO-LSSVM improves the prediction accuracyof network traffic,the predicted results can provide valuable reference information for network administrators.%针对最小二乘支持向量机参数优化问题，提出一种变异粒子群算法优化最小二乘支持向量的网络流量预测模型（MPSO-LSSVM）。首先对网络流量序列进行相空间重构，构建最小二乘支持向量的学习样本；然后采用变异粒子群算法选择最小二乘支持向量机参数，从而建立最优的网络流量预测模型，最后与其他模型进行对比实验。对比结果表明，相对于对比模型，MPSO-LSSVM提高了网络流量的预测精度，预测结果可以为网络管理员提供有价值参考信息。
展开▼
