A combined traffic flow forecasting model based on graph convolutional network and attention mechanism

Hong Zhang, Linlong Chen (), Jie Cao (), Xijun Zhang () and Sunan Kan ()
Additional contact information
Hong Zhang: College of Computer & Communication, Lanzhou University of Technology, Lanzhou 730050, P. R. China
Linlong Chen: College of Computer & Communication, Lanzhou University of Technology, Lanzhou 730050, P. R. China
Jie Cao: College of Computer & Communication, Lanzhou University of Technology, Lanzhou 730050, P. R. China
Xijun Zhang: College of Computer & Communication, Lanzhou University of Technology, Lanzhou 730050, P. R. China
Sunan Kan: College of Computer & Communication, Lanzhou University of Technology, Lanzhou 730050, P. R. China

International Journal of Modern Physics C (IJMPC), 2021, vol. 32, issue 12, 1-21

Abstract: Accurate traffic flow forecasting is a prerequisite guarantee for the realization of intelligent transportation, but due to the complex spatiotemporal characteristics of traffic flow, its forecasting has always been difficult. Deep learning can learn the deep spatiotemporal characteristics of traffic flow from a large amount of data. Deep learning can learn the deep spatiotemporal characteristics of traffic flow from a large amount of data. This paper establishes a novel combination forecasting model GGCN-SA based on deep learning for traffic flow to effectively capture the spatiotemporal characteristics of traffic flow and improve forecasting accuracy. The model captures the spatial correlation of the road traffic network through the graph convolutional network (GCN), captures the time dependence of the traffic flow through the gated recursive unit (GRU), and further introduces the soft attention mechanism (Soft Attention) to aggregate different neighborhoods Spatio-temporal information within the range to enhance the model’s ability to characterize the temporal and spatial characteristics of traffic flow. A large number of experiments have been conducted on the METR-LA and SZ-taxi data sets. The experimental results show that the GGCN-SA model proposed in this paper has better forecasting performance compared with the baseline methods.

Keywords: Traffic flow forecasting; deep learning; attention mechanism; graph convolutional network; spatiotemporal characteristics (search for similar items in EconPapers)
Date: 2021
References: Add references at CitEc
Citations:

Downloads: (external link)
http://www.worldscientific.com/doi/abs/10.1142/S0129183121501588
Access to full text is restricted to subscribers

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:wsi:ijmpcx:v:32:y:2021:i:12:n:s0129183121501588

Ordering information: This journal article can be ordered from

DOI: 10.1142/S0129183121501588

Access Statistics for this article

International Journal of Modern Physics C (IJMPC) is currently edited by H. J. Herrmann

More articles in International Journal of Modern Physics C (IJMPC) from World Scientific Publishing Co. Pte. Ltd.
Bibliographic data for series maintained by Tai Tone Lim ().