Research on traffic flow prediction based on dual-channel hybrid graph learning with trend-aware attention

Abstract Accurate traffic flow prediction is a key component of intelligent transportation systems and is crucial for improving road-network efficiency and traffic-safety management. However, existing methods still face major challenges in long-term forecasting, including insufficient spatio-tempora...

Ausführliche Beschreibung

Bibliografische Detailangaben

Hauptverfasser:	Kechao Zhu, Linlong Chen, Danfeng Wang, Kun Fan
Format:	Artigo
Sprache:	Inglês
Veröffentlicht:	Nature Portfolio 2026-04-01
Schriftenreihe:	Scientific Reports
Schlagworte:	Traffic flow prediction Spatio-temporal graph neural network Trend-aware attention Multi-scale decomposition
Online-Zugang:	https://doi.org/10.1038/s41598-026-48316-6
Tags:	Tag hinzufügen Keine Tags, Fügen Sie das erste Tag hinzu!