关键词: 短时交通流预测, SARIMA 模型, RBF 神经网络, 历史周期性, 空间相关性

Abstract:

According to the historical cycles and spatial correlation of traffic flows,a model named as SARIMA-RBF is proposed to forecast the short- term traffic flow,which integrates the advantage of the SARIMA model inmaking use of the historical cyclic data with that of the RBF model in making use of the spatial correlation data.Inthe proposed model,the SARIMA model is adopted to forecast the traffic flow at the next time by using historicaldata,and then the RBF model is employed to obtain the output value by combining the SARIMA- based forecastdata with the relevant traffic flow data of the upstream and downstream of test point.The output value of the RBFmodel is exactly the prediction result of the SARIMA- RBF model.Experimental results show that,in comparisonwith the SARIMA model and the RBF model,the SARIMA- RBF model achieves better results in forecasting theshort- term traffic flow,because it considers both the historical cycles and the spatial correlation.

Key words: short- term traffic flow forecasting, SARIMA model, RBF networks, historical cycle, spatial correlation