SGGformer: Shifted Graph Convolutional Graph-Transformer for Traffic Prediction.

Area of Science:

• Intelligent Transportation Systems
• Data Science
• Machine Learning

Background:

• Accurate traffic flow prediction is crucial for the safe and stable development of intelligent cities.
• Complex spatiotemporal correlations within traffic data present significant challenges for existing prediction models.

Purpose of the Study:

• To propose SGGformer, an advanced traffic grade prediction model designed to overcome the limitations of current methods.
• To enhance the accuracy and efficiency of traffic flow prediction in intelligent urban environments.

Main Methods:

• Utilized a shifted window operation for time series data coarsening to reduce computational complexity.
• Employed a multi-channel graph convolutional network to capture and aggregate multi-dimensional spatial road correlations.
• Developed an improved graph Transformer network to effectively extract long-term temporal correlations from traffic data.

Main Results:

• The SGGformer model demonstrated superior prediction performance compared to state-of-the-art baselines.
• Empirical evaluation using actual traffic datasets validated the model's effectiveness.

Conclusions:

• SGGformer offers a robust solution for accurate traffic prediction in intelligent cities.
• The integration of novel techniques in SGGformer significantly advances the field of spatiotemporal traffic forecasting.