Trajectory-User Linking via Multi-Scale Graph Attention Network
This work has been accepted by the journal "Pattern Recognition"
Environment：
Numpy 1.24.2
PyTorch 1.13.1
torch-geometric 2.2.0
Baselines：
TULER: Identifying Human Mobility via Trajectory Embeddings(https://www.ijcai.org/Proceedings/2017/0234.pdf)(IJCAI2017)
TULVAE: Trajectory-User Linking via Variational AutoEncoder(https://www.ijcai.org/Proceedings/2018/0446.pdf)(IJCAI2018)
TULAR: Trajectory-User Link with Attention Recurrent Networks(https://ailb-web.ing.unimore.it/icpr/media/posters/11413.pdf)(ICPR 2020)
STULIG: Toward Discriminating and Synthesizing Motion Traces Using Deep Probabilistic Generative Models(https://ieeexplore.ieee.org/document/9165954)(TNNLS 2021)
GNNTUL: Trajectory-User Linking via Graph Neural Network(https://ieeexplore.ieee.org/document/9500836)(ICC 2021)
TULRN: TULRN: Trajectory user linking on road networks(https://link.springer.com/article/10.1007/s11280-022-01124-0)(WWWJ 2023)