BNs-OLS-SARIMA对城市短时交通流的预测

Urban short-term traffic flow forecasting studied by BNs-OLS-SARIMA model

通过在目标路口构建贝叶斯交通网(BNs),并对与此交通网相关的交通流建立非平稳季节(SARIMA)模型,采用最小二乘法(OLS)取得相应模型的最佳权重组合,对缺失数据下的城市道路短时交通流进行预测。使用重庆市某路口的交通流数据对模型进行检测,通过多种预测指标对结果进行对比分析,结果表明BNs-OLS-SARIMA把交通流的网络结构与其周期性结合在一起,对短时交通流有良好的预测效果。

This paper analyzed the proplem of incomplete data’s short-term urban road traffic flow prediction by building Bayesian networks （ BNs） of the traffic structure on the target intersection,and modeling the related traffic flows with seasonal autoregressive integrated moving average （ SARIMA） models calculated whose weight by the least squares （ OLS） . Data of Chongqing’s traffic flow tested predictive power of the model,and compared the method’s result by a few indicators. The analysis of the results considers both the network structure and periodicity and show the BNs-OLS-SARIMA model has a better effect on short-term traffic flow prediction.