Geo-aware graph-augmented self-attention network for individual mobility prediction

Yahui Wang, Hongchang Chen, Shuxin Liu, Kai Wang, Yuxiang Hu

Abstract: Highlights•A heterogeneous location graph is constructed.•A simple and efficient learning method is used to learn the location graph.•The impacts of geospatial distance and location transition have coupling effects.•Multi-scale time encoding is designed.