MGSFformer: A Multi-Granularity Spatiotemporal Fusion Transformer for air quality prediction

Chengqing Yu, Fei Wang, Yilun Wang, Zezhi Shao, Tao Sun, Di Yao, Yongjun Xu

Abstract: Highlights•A new Transformer is proposed for air quality spatiotemporal prediction.•MGSFformer full fuses multi-granularity and spatiotemporal correlations of data.•The proposed model can outperform 11 SOTA models on 3 datasets.