A Data Grouping CNN Algorithm for Short-Term Traffic Flow Forecasting

摘要：

　　In this paper,a data grouping approach based on convolutional neural network (DGCNN) is proposed for forecasting urban shortterm traffic flow.This approach includes the consideration of spatial relations between traffic locations,and utilizes such information to train a convolutional neural network for forecasting.There are three advantages of our approach: (1) the spatial relations of traffic flow are adopted;(2) high-quality features are extracted by CNN;and (3) the accuracy of forecasting short-term traffic flow is improved.To verify our model,extensive experiments are performed on a real data set,and the result shows that the model is more effective than other existing methods.

关键词： Convolution Neural Network Traffic flow forecasting CBOW Deep learning

作者: Donghai Yu Yang Liu Xiaohui Yu

作者单位: School of Computer Science and Technology,Shandong University,Jinan 250101,China School of Computer Science and Technology,Shandong University,Jinan 250101,China;School of Informati

会议类型: 国际会议

会议名称: International Asia-Pacific Web Conference(第18届国际亚太互联网大会)

会议地点: 苏州

会议语种:英文

页码: 92-103

在线出版日期: 2016-09-23（万方平台首次上网日期，不代表论文的发表时间）