Tag
Proposes a Global-Local Graph Attention Network (GLGAT) with pairwise encoding and event-based adjacency matrix for traffic forecasting, effectively capturing spatio-temporal correlations and achieving competitive performance on real-world datasets.