The SAE-NAD model for Point-of-Interest Recommendation

The implementation of the paper:

Chen Ma, Yingxue Zhang, Qinglong Wang, and Xue Liu, "Point-of-Interest Recommendation: Exploiting Self-Attentive Autoencoders with Neighbor-Aware Influence", in the 27th ACM International Conference on Information and Knowledge Management (CIKM 2018)

Arxiv: https://arxiv.org/abs/1809.10770

Please cite our paper if you use our code. Thanks!

Author: Chen Ma ([email protected])

Bibtex

@inproceedings{DBLP:conf/cikm/MaZWL18,
  author    = {Chen Ma and
               Yingxue Zhang and
               Qinglong Wang and
               Xue Liu},
  title     = {Point-of-Interest Recommendation: Exploiting Self-Attentive Autoencoders with Neighbor-Aware Influence},
  booktitle = {{CIKM}},
  pages     = {697--706},
  publisher = {{ACM}},
  year      = {2018}
}

Environments

• python 3.6
• PyTorch (version: 0.4.0)
• numpy (version: 1.15.0)
• scipy (version: 1.1.0)
• sklearn (version: 0.19.1)

Dataset

In our experiments, the Foursquare and Yelp datasets are from http://spatialkeyword.sce.ntu.edu.sg/eval-vldb17/. And the Gowalla dataset is from https://snap.stanford.edu/data/loc-gowalla.html (if you need the data after preprocessing, please send me an email).

Example to run the code

Data preprocessing:

Run the cal_poi_pairwise_relation.py to calculate the pairwise relations between locations, which will be stored in ./data/Foursquare/.

python cal_poi_pairwise_relation.py	

Train and evaluate the model (you are strongly recommended to run the program on a machine with GPU):

python run.py