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

TLDR

Accurate traffic forecasting is essential for safety, stability, and efficiency of intelligent transportation systems, yet it remains challenging due to the need to model temporal and spatial dynamics, capture periodicity and spatial heterogeneity, and perform well on long‑term predictions. This study introduces an Attention‑based Spatial‑Temporal Graph Neural Network (ASTGNN) to address these forecasting challenges. ASTGNN employs a temporal self‑attention module that leverages local context for sequence transformation, a dynamic graph convolution module that uses self‑attention to capture spatial correlations, and embedding modules that explicitly model periodicity and spatial heterogeneity. Experiments on five real‑world traffic flow datasets demonstrate that ASTGNN surpasses state‑of‑the‑art baselines.

Abstract

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.