一种短时交通流组合预测模型

为提高短时交通流预测的精度,提出利用BP神经网络、RBF神经网络和ARIMA模型构建组合预测模型,该组合预测模型利用最优化原理进行权系数的分配,并且满足分配到的权值始终具有实际意义。通过对分配的权系数进行显著性检验,以确保组合预测模型中选用的单项预测方法显著相关。通过实例分析,验证了组合预测模型的有效性,结果表明,相比较单一的预测模型,组合预测模型具有更高的预测精度。

Combination Prediction of Short-time Traffic Flow Based on Regression

In order to improve the accuracy of short-term traffic flow prediction,a forecasting modelwas proposed based on the combination of BP neural network,RBF neural network and ARIMA model.This proposed model determined the weights through the optimization principle and ensured that the assigned weight always had practical significance;and a significant test was performed for the determined weights for ensuring that the selected prediction methods were significantly correlated.The real traffic flow data was used to test the proposed combined model,and the effectiveness of the combined forecasting model is verified.The results also showed that the performance of combined forecasting model out performed the single prediction model,which had higher prediction accuracy than the single prediction model.