AGCL [TKDE 2024]

This is the implementation for the paper: “Next Point-of-Interest Recommendation with Adaptive Graph Contrastive Learning.”

Preliminaries

Conda Environment

  conda env create -f environment.yml

Requirements

• Python 3.9.17
• pytorch 1.10.0
• pandas 2.0.3
• numpy 1.25.1
• setuptools 59.5.0 -> pip install setuptools==59.5.0
• torch-summary 1.4.5 -> pip install torch-summary

  pip install torch==1.10.0+cu111 torchvision==0.11.0+cu111 torchaudio==0.10.0 -f https://download.pytorch.org/whl/torch_stable.html

Datasets

We use SIN, NYC and Gowalla datasets. The processed SIN and Gowalla datasets are from ARGAN, and we preprocess NYC dataset by data_process.py. For more details of data preprocessing, please refer to our paper or data_process.py:

python data_process.py

Graph

We use Knowledge Graph Embedding (KGE) to construct graphs. The processed code are from Graph-Flashback. For more details of graph construction, please refer to KGE directory:

python data_process.py

Model Training

To train our model with default hyper-parameters:

python main.py

Acknowledgement

The code is implemented based on ARGAN.

Citing

If you use AGCL in your research, please cite the following paper:

@article{DBLP:journals/tkde/RaoJSCHYK25,
  author       = {Xuan Rao and
                  Renhe Jiang and
                  Shuo Shang and
                  Lisi Chen and
                  Peng Han and
                  Bin Yao and
                  Panos Kalnis},
  title        = {Next Point-of-Interest Recommendation With Adaptive Graph Contrastive
                  Learning},
  journal      = {{IEEE} Trans. Knowl. Data Eng.},
  pages        = {1366--1379},
  year         = {2025}
}