Short-Term Traffic Flow Combination Forecast by Co-integration Theory

Short-term traffic flow prediction is a priority issue of intelligent transportation system. The accuracy of the prediction results directly affects the traffic control and management. Therefore, it is the key technology for the advanced traffic management information system. This paper briefly describes the concept of co-integration and error correction model, and then verifies the validity of the combination of traffic flow forecasting model using the co-integration of series. It also improves the stability of the combination forecasting model through the error correction model. The historical and real-time traffic flow data, collected form the Second Ring Road of Beijing, are used to verify the model. The results indicate that the combination model based on the co-integration and error correction model meets the actual traffic flow characteristics well and obtains a better prediction result.

摘要交通流短时预测是智能交通系统中的一个重点问题，预测效果的好坏直接关系到控制和诱导的结果，是实现先进交通管理信息系统的关键技术之一. 本文简要介绍了协整和误差修正模型的概念，利用序列的协整性来进行交通流组合预测模型的有效性验证，并利用误差修正模型提高组合预测模型的稳定性. 我们利用北京市二环路上采集到的交通流数据进行了模型的验证. 研究结果表明，基于协整理论的交通流组合预测模型可以取得很好的预测效果.