Journal of Computer Applications ›› 2020, Vol. 40 ›› Issue (5): 1488-1493.DOI: 10.11772/j.issn.1001-9081.2019091568

• Frontier & interdisciplinary applications • Previous Articles     Next Articles

Urban road short-term traffic flow prediction based on spatio-temporal node selection and deep learning

CAO Yu1, WANG Cheng1, WANG Xin1, GAO Yueer2   

  1. 1.College of Computer Science and Technology, Huaqiao University, XiamenFujian 361021, China
    2.School of Architecture, Huaqiao University, XiamenFujian 361021, China
  • Received:2019-09-16 Revised:2019-11-20 Online:2020-05-10 Published:2020-05-15
  • Contact: WANG Cheng, born in 1984, Ph. D., associate professor. His research interests include traffic big data, data mining, machine learning.
  • About author:CAO Yu, born in 1995, M. S. candidate. Her research interests include traffic big data.WANG Cheng, born in 1984, Ph. D., associate professor. His research interests include traffic big data, data mining, machine learning.WANG Xin, born in 1995, M. S. candidate. His research interests include data mining, signal processing.GAO Yueer, born in 1983, Ph. D., associate professor. Her research interests include traffic big data, data mining.
  • Supported by:

    This work is partially supported by the Youth Program of the National Natural Science Foundation of China (51608209),the Surface Program of Natural Science Foundation of Fujian Province (2017J01090), the Project of Guiding Plan of Fujian Province (2019H0017), the Science and Technology Program of Quanzhou (2018Z008), the Postgraduate Research and Innovation Ability Cultivation Project of Huaqiao University (17014083001).


曹堉1, 王成1, 王鑫1, 高悦尔2   

  1. 1.华侨大学 计算机科学与技术学院,福建厦门 361021
    2.华侨大学 建筑学院,福建厦门 361021
  • 通讯作者: 王成(1984—)
  • 作者简介:曹堉(1995—),女,福建龙岩人,硕士研究生,CCF会员,主要研究方向:交通大数据; 王成(1984—),男,湖北咸宁人,副教授,博士,CCF高级会员,主要研究方向:交通大数据、数据挖掘、机器学习; 王鑫(1995—),男,山西太原人,硕士研究生,主要研究方向:数据挖掘、信号处理; 高悦尔(1983-),女,福建泉州人,副教授,博士,主要研究方向:交通大数据、数据挖掘。
  • 基金资助:

    国家自然科学基金青年基金资助项目(51608209);福建省自然科学基金面上项目(2017J01090);福建省引导性计划项目 (2019H0017);泉州市科技计划项目(2018Z008);华侨大学研究生科研创新能力培育计划项目(17014083001)。


In order to solve the problems of insufficient consideration of the traffic flow characteristics and the low accuracy of the prediction, a short-term prediction method of urban road traffic flow based on spatio-temporal node selection and deep learning was proposed. Firstly, the characteristics of traffic flow were analyzed in theory and data representation to obtain its spatial characteristics, and temporal characteristics and candidate spatio-temporal nodes set. Secondly, the set of candidate spatio-temporal nodes was determined according to the reachable range of traffic flow, and the fitness was calculated by taking the inverse of the sum of squares of errors as the objective function. In the historical training set, genetic algorithm and Back Propagation Neural Network (BPNN) were used to select spatio-temporal nodes, and the final spatio-temporal nodes and BPNN structure were obtained. Finally, the measured values of the selected spatio-temporal nodes were taken as the input of BPNN in the working set to obtain the predicted values. The experimental results show that compared with only using data of adjacent spatio-temporal nodes, using other time node ranges, Support Vector Machine (SVM) and Gradient Boosting Decision Tree (GBDT), the proposed model has a slight reduction in Mean Absolute Error (MAE) and Mean Absolute Percentage Error (MAPE), which are 10.631 6 and 14.275 8%, respectively; and 0.257 3和0.999 1 percentage points lower than those by using adjacent spatio-temporal nodes.

Key words: urban traffic, short-term traffic flow prediction, spatio-temporal correlation analysis, spatio-temporal node selection, deep learning, wrapped feature selection, Genetic Algorithm (GA)


针对目前交通流短时预测对于交通流特性考虑得不够全面、预测精度不高的问题,提出一种基于时空节点选择和深度学习的城市道路交通流短时预测方法。首先,在理论和数据表现上对交通流特性进行分析,获得时空特性;其次,根据车流的可达范围确定候选时空节点集合,以误差平方和的倒数为目标函数计算适应度,在训练集上使用遗传算法和反向传播神经网络(BPNN)进行时空节点选择,得到最终的时空节点和训练好的BPNN;最后,在工作集上将选择的时空节点的实测值输入训练好的BPNN得出预测值。实验结果表明,所提模型与仅使用相邻时空节点数据、采用其他时间节点范围、支持向量机(SVM)和梯度提升树(GBDT)相比误差略有降低,平均绝对误差(MAE)和平均绝对百分误差(MAPE)分别为10.631 6和14.275 8%;仅使用与待预测路段相邻空间的交通流数据的预测结果相比MAE和MAPE两个值上分别高出了0.257 3和0.999 1个百分点。

关键词: 城市交通, 短时交通流预测, 时空相关性分析, 时空节点选择, 深度学习, 包裹式特征选择, 遗传算法

CLC Number: