Code of paper "Temporal Meta-path Guided Explainable Recommendation".

python==3.6.12
networkx==2.5
numpy==1.15.0
pandas==1.0.1
pytorch==1.0.0
pytorch-nlp==0.5.0
gensim==3.8.3

You can also install the environment via requirements.txt and environment.yaml.

The original data can be found in the amazon data website.

For example, the meta_Musical_Instruments.json of Amazon_Music can be found here. The user_rate_item.csv in the code is here (ratings only).

If you want to change the dataset, you can modify the name in the code.

1.process data (You can ignore this step, if you just want to check TMER.)

python data_process.py

2.learn the user and item representations

python data/path/embed_nodes.py

3.learn the item-item path representations

python data/path/user_history/item_item_representation.py

4.learn the user-item path representations

python data/user_item_representation.py

5.generate user-item and item-item meta-path instances and learn their representations

python data/path/generate_paths.py
python data/path/user_history/meta_path_instances_representation.py

6.sequence item-item paths for each user

python data/path/user_history/user_history.py

7.run the recommendation

python run.py

If you find this code useful in your research, please consider citing:

@article{chen2021temporal,
  title={Temporal Meta-path Guided Explainable Recommendation},
  author={Chen, Hongxu and Li, Yicong and Sun, Xiangguo and Xu, Guandong and Yin, Hongzhi},
  journal={arXiv preprint arXiv:2101.01433},
  year={2021}
}

or

@inproceedings{10.1145/3437963.3441762,
	author = {Chen, Hongxu and Li, Yicong and Sun, Xiangguo and Xu, Guandong and Yin, Hongzhi},
	title = {Temporal Meta-Path Guided Explainable Recommendation},
	year = {2021},
	booktitle = {Proceedings of the 14th ACM International Conference on Web Search and Data Mining},
	pages = {1056–1064}
}