Multi-graph Spatio-temporal Graph Convolutional Network for Traffic Flow Prediction

Weilong Ding,Tianpu Zhang,Jianwu Wang,Zhuofeng Zhao

DOI: https://doi.org/10.48550/arXiv.2308.05601

2023-08-10

Abstract:Inter-city highway transportation is significant for urban life. As one of the key functions in intelligent transportation system (ITS), traffic evaluation always plays significant role nowadays, and daily traffic flow prediction still faces challenges at network-wide toll stations. On the one hand, the data imbalance in practice among various locations deteriorates the performance of prediction. On the other hand, complex correlative spatio-temporal factors cannot be comprehensively employed in long-term duration. In this paper, a prediction method is proposed for daily traffic flow in highway domain through spatio-temporal deep learning. In our method, data normalization strategy is used to deal with data imbalance, due to long-tail distribution of traffic flow at network-wide toll stations. And then, based on graph convolutional network, we construct networks in distinct semantics to capture spatio-temporal features. Beside that, meteorology and calendar features are used by our model in the full connection stage to extra external characteristics of traffic flow. By extensive experiments and case studies in one Chinese provincial highway, our method shows clear improvement in predictive accuracy than baselines and practical benefits in business.

Machine Learning,Artificial Intelligence

What problem does this paper attempt to address?

The paper attempts to address the problem of long-term multi-location traffic flow prediction across highway networks. Specifically, the study targets the following major challenges:

1. **Data Imbalance**: In practical applications, the distribution of traffic flow data at different locations is extremely uneven, leading to a decline in prediction performance.
2. **Complex Spatiotemporal Correlation**: Existing methods fail to fully capture the complex spatiotemporal correlation features.
3. **Insufficient Consideration of External Factors**: Factors affecting traffic flow, such as weather conditions and holidays, have not been fully utilized.

To address these issues, the authors propose a method based on Multi-Graph Spatiotemporal Graph Convolutional Networks (MSTGCN). This method improves prediction performance through the following steps:
- **Data Normalization Strategy**: Handles long-tail distributed data to avoid treating extreme values at certain key sites as outliers.
- **Multi-Perspective Graph Construction**: Constructs graph models from three different perspectives—geographical, influence, and resilience relationships—to capture diverse spatial features.
- **Integration of External Factors**: Introduces calendar and weather features in the fully connected layer to improve prediction accuracy.

Through extensive experiments and case studies on a provincial highway network in China, this method shows a significant improvement in prediction accuracy compared to baseline models and has practical application value.