基于组合模型的网络流量预测

Network Traffic Prediction by Using Combination Model

网络流量是多种因素综合作用的结果,具有一定的混沌性和非线性,为了对网络流量将来变化趋势进行准确预测,提出了一种网络流量的组合预测模型(PHR-ELM)。分别采用C-C算法和Cao算法对原始网络流量数据进行混沌分析,构建极限学习机的训练样本集和测试样本集,然后将训练样本输入到极限学习机进行学习,建立网络流量预测模型,最后测试样本集输入到模型中进行单步和多步预测的验证性实验。结果表明,PHR-ELM可以反映网络流量的混沌性和非线性变化趋势,获得了高精度的网络流量预测结果,且预测结果要优于其它模型。

Network traffic is the result of many factors,which is chaos and nonlinear.A combined prediction model of network traffic is proposed to predict the future trend of network traffic accurately by using combination model.C-C algorithm and Cao algorithm are separately used for chaotic analysis of original network traffic data to construct training set and test set,and then training samples are input into extreme learning machine to learn and establish prediction model,finally,test samples are input into the model to verify the performance by using single step and multi-step prediction experiments.The results show that PHR-ELM can reflect the network traffic chaos and nonlinear trend,predict the results of network traffic with high accuracy,and the prediction results are better than other models.