Attention-based dynamic spatial-temporal graph convolutional networks for traffic speed forecasting

In recent years, spatial-temporal graph modeling based on graph convolutional neural networks (GCN) has become an effective method for mining spatial-temporal dependencies in traffic forecasting research. However, existing studies lack the capability of dynamic spatial-temporal modeling of traffic speeds. Furthermore, longterm forecasting is difficult because of the diversity of traffic conditions. In addition, traditional studies capture only the features of fixed graph structures, which do not reflect real spatial dependence. To address these challenges, this study proposes a novel attention-based dynamic spatial-temporal graph convolutional network (ADSTGCN) model. ADSTGCN mainly consists of multiple dynamic spatial-temporal blocks, each of which contains three modules: 1) a dynamic adjustment module to model the dynamic spatial-temporal correlations of traffic speed, 2) a gated dilated convolution module to mine long-term dependencies, and 3) a spatial convolution module to capture hidden spatial dependencies. Experiments on three public traffic datasets demonstrated the good performance of the model.

Automated summary

This study proposes an attention-based dynamic spatial-temporal graph convolutional network (ADSTGCN) model to address the challenges of spatial-temporal modeling and long-term forecasting in traffic forecasting research. The model consists of multiple dynamic spatial-temporal blocks, each containing modules for dynamic adjustment, gated dilated convolution, and spatial convolution. Experimental results on three public traffic datasets demonstrate the model's good performance.