Learning to Rank Rationales for Explainable Recommendation

This is the official code implementation for arXiv manuscript: Learning to Rank Rationales for Explainable Recommendation

Citation:

@article{xu2022learning,
  title={Learning to Rank Rationales for Explainable Recommendation},
  author={Xu, Zhichao and Han, Yi and Yang, Tao and Tran, Anh and Ai, Qingyao},
  journal={arXiv preprint arXiv:2206.05368},
  year={2022}
}

Dependency

pip3 install -r requirements.txt

Download the EXTRA dataset from Link

put IDs.pickle, id2exp.json, train.index and test.index under same folder

Preprocess

python preprocess.py --input_dir your_input_dir

Build user, item and doc representations

python build_reps.py --input_dir your_input_dir

Train SE-BPER

python train.py --input_dir your_input_dir