TA-STGCN: trend-aware attention spatio-temporal graph convolutional network for traffic flow prediction

Key Points

• Predictions using the model show an increase in accuracy compared to traditional methods, particularly in dynamic environments.
• A notable improvement of 15% in prediction accuracy was observed during peak traffic hours, emphasizing real-world relevance.
• Analysis utilizes a spatio-temporal graph convolutional network based on real-time traffic data from multiple city intersections.
• The model's innovative design may enable more efficient traffic management systems in urban areas.