针对网络流量的预测问题,结合网络流量序列的自相似性分析而提出一种基于差分自回归滑动平均模型(ARIMA)补偿极限学习机(ELM)的网络流量预测方法.首先利用ELM模型对网络流量序列进行一步预测,然后对网络流量预测的误差序列通过ARIMA模型进行修正,最后将ELM模型预测值与ARIMA模型修正值进行叠加得到最终的预测值。与单独的ARIMA模型、最小二乘支持向量机(LS—SVM)预测模型以及Elman神经网络预测模型进行了对比,仿真结果表明本文的方法具有更高的预测精度。
For the problem of network traffic prediction, we propose a network traffic prediction method based on an autoregressive integrated moving average model (ARIMA) compensation extreme learning machine (ELM), combining the network traffic sequence self-similar analysis. Firstly, one-step network traffic is precJictecl by the extreme learning machine, and then the sequence of network traffic forecast error is corrected by the ARI- MA model. Finally, we obtain the network traffic by adding the predictive values of ELM and ARIMA. We compare the proposed ARIMA method with least squares support vector machine and Elman neural network pre- diction model. Simulation results show that the proposed method has higher prediction accuracy than the either.