基于ARIMA和小波神经网络组合模型的交通流预测

摘要

Aimed at the limitation of low prediction accuracy at the present stage of city road traffic,a prediction method is proposed based on Hybrid Autoregressive Integrated Moving Average ( ARIMA) and Wavelet Neural Network ( WNN) to predict traffic flow. Using the good linear fitting ability of ARIMA and the strong nonlinear mapping ability of WNN,the traffic flow time series are considered to be composed of a linear autocorrelation structure and a nonlinear structure. ARIMA model is used to predict the linear component of traffic flow time series and the wavelet neural network model is applied to the nonlinear residual component prediction. The simulation results show that the hybrid model can produce more accurate prediction than that of single model,which improves prediction accuracy of traffic flow prediction,and it’ s an efficient method.%针对现阶段城市道路交通流预测精度不高的局限性，提出了一种基于差分自回归滑动平均( ARIMA)和小波神经网络( WNN)组合模型的预测方法来进行交通流预测。利用差分自回归滑动平均模型良好的线性拟合能力和小波神经网络模型强大的非线性关系映射能力，把交通流时间序列的数据结构分解为线性自相关结构和非线性结构两部分。采用差分自回归滑动平均模型预测交通流序列的线性部分，用小波神经网络模型预测其非线性残差部分，最终合成为整个交通流序列的预测结果。计算机仿真结果表明：组合模型的预测精度高于ARIMA模型和WNN模型各自单独使用时的预测精度，组合模型可以提高交通流预测精度，是交通流预测的有效方法。