Attention Based Spatial-Temporal Graph Convolutional Networks for Traffic Flow Forecasting
Графовые сверточные сети с пространственно-временным вниманием для прогнозирования транспортных потоков
2019-07-17
SCID: 54.1/g7tyxp58
Discuss with AI
Caltrans Performance Measurement System (PeMS)graph convolutionsspatial-temporal attention mechanismspatial-temporal graph convolutional networkstraffic flow forecasting
Figures from the paper
Abstract (AI)
Forecasting the traffic flows is a critical issue for researchers and practitioners in the field of transportation. However, it is very challenging since the traffic flows usually show high nonlinearities and complex patterns. Most existing traffic flow prediction methods, lacking abilities of modeling the dynamic spatial-temporal correlations of traffic data, thus cannot yield satisfactory prediction results. In this paper, we propose a novel attention based spatial-temporal graph convolutional network (ASTGCN) model to solve traffic flow forecasting problem. ASTGCN mainly consists of three independent components to respectively model three temporal properties of traffic flows, i.e., recent, daily-periodic and weekly-periodic dependencies. More specifically, each component contains two major parts: 1) the spatial-temporal attention mechanism to effectively capture the dynamic spatialtemporal correlations in traffic data; 2) the spatial-temporal convolution which simultaneously employs graph convolutions to capture the spatial patterns and common standard convolutions to describe the temporal features. The output of the three components are weighted fused to generate the final prediction results. Experiments on two real-world datasets from the Caltrans Performance Measurement System (PeMS) demonstrate that the proposed ASTGCN model outperforms the state-of-the-art baselines.
Key Findings
1
ASTGCN forecasts traffic flow by modeling dynamic spatial-temporal correlations with an attention-based graph convolutional architecture.
2
Each component combines spatial-temporal attention with graph convolutions for spatial patterns and standard convolutions for temporal features.
3
Experiments on two real-world Caltrans PeMS datasets show that ASTGCN outperforms state-of-the-art baseline methods.
4
Outputs from the three temporal components are weighted and fused to produce final traffic-flow predictions.
5
The model uses three independent components to capture recent, daily-periodic, and weekly-periodic traffic dependencies.
Research Object
Traffic flow time series on road sensor/network (spatiotemporal traffic data)
Research Subject
their nonlinear spatiotemporal patterns and dynamic spatial-temporal correlations across recent, daily-periodic, and weekly-periodic dependencies, for forecasting
Publication Details
Publication Date
2019-07-17
Journal
Publisher
ISSN
Cited by
4142
Open access PDF
Access Type
Author Information
Download PDF
Subscribe to digest
References available in scid.ai5
Deep Residual Learning for Image Recognition2016
Semi-Supervised Classification with Graph Convolutional Networks2016
Empirical Evaluation of Gated Recurrent Neural Networks on Sequence Modeling2014
Spatio-Temporal Graph Convolutional Networks: A Deep Learning Framework for Traffic Forecasting2018
Detecting Functionality-Specific Vulnerabilities via Retrieving Individual Functionality-Equivalent APIs in Open-Source Repositories2025
Cited by9
A Comprehensive Survey on Graph Neural Networks2020
Graph neural networks: A review of methods and applications2020
Graph Neural Networks: A Review of Methods and Applications2018
Spatio-Temporal Graph Neural Networks for Predictive Learning in Urban Computing: A Survey2023
A Survey on Graph Neural Networks for Time Series: Forecasting, Classification, Imputation, and Anomaly Detection2024
Decoupled dynamic spatial-temporal graph neural network for traffic forecasting2022
Graph Neural Networks for Intelligent Transportation Systems: A Survey2023
Computing Graph Neural Networks: A Survey from Algorithms to Accelerators2021
A Survey on Graph Representation Learning Methods2023