基于Elman神经网络改进ARIMA模型的网络负荷预测研究

摘要

针对ARIMA模型预测结果具有误差大和滞后性的缺点,利用Elman神经网络的优越的非线性逼近能力修正ARIMA模型,提出一种基于ARIMA-Elman组合模型的网络负荷预测模型.选择RMSE、MAE和MAPE作为评价指标,研究结果表明,提出的ARIMA-Elman组合模型进行网络负荷数据预测的精度优于ARIMA、SVM和Elman,从而验证了算法的优越性和可靠性.%For the prediction results of ARIMA model has large error and the disadvantage of lag,using the nonlinear approximation capability of the Elman neural network is superior modified ARIMA model,put forward a kind of network load forecasting model based on ARIMA-Elman's combination model.RMSE,MAE and MAPE is chosen as the evaluation index,the results show that the proposed Elman combination of ARIMA model for network load data to predict the accuracy is better than ARIMA,SVM and Elman,which verified the superiority and reliability of the algorithm in this paper.