MMPOI

This is the PyTorch implementation of the paper "MMPOI: A Multi-Modal Content-Aware Framework for POI Recommendations"

Train

• Download dataset from https://drive.google.com/file/d/17FbNvkO74xub6AeT2fpm938qDqUDB-04/view?usp=sharing.

• Unzip NYC.zip to dataset/.

• Run build_graph.py to construct the sequence graphs from the training data.

• Train the model using python train.py. All hyper-parameters are defined in param_parser.py

Raw Dataset

NYC: https://drive.google.com/file/d/1v0BvKs46ixUf1CgjRlk9MY5JsFL7ILxC/view?usp=sharing

TKY: https://drive.google.com/file/d/1Cpnp2iEmHGfvUkOL-8myeQmyOuq4LmV7/view?usp=sharing

Please cite our paper if you use these datasets, thank you very much!

Citation

@inproceedings{DBLP:conf/www/XuCZC24,
  author       = {Yang Xu and
                  Gao Cong and
                  Lei Zhu and
                  Lizhen Cui},
  title        = {{MMPOI:} {A} Multi-Modal Content-Aware Framework for {POI} Recommendations},
  booktitle    = {Proceedings of the {ACM} on Web Conference 2024, {WWW} 2024, Singapore,
                  May 13-17, 2024},
  pages        = {3454--3463},
  publisher    = {{ACM}},
  year         = {2024},
  url          = {https://doi.org/10.1145/3589334.3645449},
  doi          = {10.1145/3589334.3645449}
}