Learning Dynamics and Heterogeneity of Spatial-Temporal Graph Data for Traffic Forecasting

By Shengna Guo, Youfang Lin, Huaiyu Wan, Xiucheng Li, Gao Cong. IEEE Transactions on Knowledge and Data Engineering, 2021

718 citations recorded by the source corpus.

Accurate traffic forecasting is critical in improving safety, stability, and efficiency of intelligent transportation systems. Despite years of studies, accurate traffic prediction still faces the following challenges, including modeling the dynamics of traffic data along both temporal and spatial dimensions, and capturing the periodicity and the spatial heterogeneity of traffic data, and the problem is more difficult for long-term forecast. In this paper, we propose an Attention based Spatial-Temporal Graph Neural Network (ASTGNN) for traffic forecasting. Specifically, in the temporal dimension, we design a novel self-attention mechanism that is capable of utilizing the local context, which is specialized for numerical sequence representation transformation. It enables our prediction model to capture the temporal dynamics of traffic data and to enjoy global receptive fields that is beneficial for long-term forecast. In the spatial dimension, we develop a dynamic graph convolution module, employing self-attention to capture the spatial correlations in a dynamic manner. Furthermore, we explicitly model the periodicity and capture the spatial heterogeneity through embedding modules. Experiments on five real-world traffic flow datasets demonstrate that ASTGNN outperforms the state-of-the-art baselines.

Subfields: Building and Construction, Signal Processing

View Paper: https://www.semanticscholar.org/paper/fbaaffc699b17256f1269e6b27d65741fbcab840

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Building and Construction

1. Junbo Zhang
2. Yu Zheng
3. Yuxuan Liang
4. Jiawei Jiang
5. Zezhi Shao
6. Jingyuan Wang
7. Fei Wang
8. Zhao Zhang

In Signal Processing

1. Junbo Zhang
2. Jiawei Jiang
3. Zezhi Shao
4. Jingyuan Wang
5. Chengkai Han
6. Yu Zheng
7. Wayne Xin Zhao
8. Xing Xie

Citations in the corpus

• Spatio-Temporal Graph Neural Networks for Predictive Learning in Urban Computing: A Survey (2023, background): “The object of network-wide prediction is usually the traffic flow or speed on the given road networks [15], [18], [24]–[26].”
• PDFormer: Propagation Delay-aware Dynamic Long-range Transformer for Traffic Flow Prediction (2023, background): “Recently, the attention mechanism has become increasingly popular due to its effectiveness in modeling the dynamic dependencies in traffic data (Guo et al. 2019; Wang et al. 2020; Lin et al. 2020b; Guo et al. 2021; Ye et al. 2022).”
• Decoupled Dynamic Spatial-Temporal Graph Neural Network for Traffic Forecasting (2022, neutral): “The attention mechanism [35] and its variants, which theoretically have infinite receptive field size, are widely used to capture long-term temporal dependencies [12] in the sequential model [37, 53].”
• Pre-training Enhanced Spatial-temporal Graph Neural Network for Multivariate Time Series Forecasting (2022, methodology): “For PEMS04, we use about 60% of data for training, 20% of data for testing, and the remaining 20% for validation [11, 12].”