Trend-aware spatio-temporal fusion graph convolutional network with self-attention for traffic prediction

Xiongtao Zhang, Lijie Pan, Qing Shen, Zhenfang Liu, Jungang Lou, Yunliang Jiang

Abstract: Highlights•Multilevel road graph is integrated to enhance prediction accuracy.•The Temporal Trend-Aware module improves long-term temporal learning.•Trend-Aware addresses uneven data distribution and improves node matching.•Training and Inference speed is significantly improved.