bper's Introduction

BPER+ (Bayesian Personalized Explanation Ranking enhanced by BERT)

Papers

Lei Li, Yongfeng Zhang, Li Chen. On the Relationship between Explanation and Recommendation: Learning to Rank Explanations for Improved Performance. ACM Transactions on Intelligent Systems and Technology (TIST), 2022.
Lei Li, Yongfeng Zhang, Li Chen. EXTRA: Explanation Ranking Datasets for Explainable Recommendation. SIGIR'21 Resource.

Datasets to download

Amazon Movies & TV
TripAdvisor Hong Kong
Yelp 2019

If you are interested in how to create the datasets, please refer to EXTRA.

Usage

Below is an example of how to run BPER+.

python -u run_bperp.py \
--cuda \
--data_dir ../Amazon/ \
--index_dir ../Amazon/2/ \
--lr 0.0001 >> bperp.log

Use pre-downloaded BERT

Download the three files and put them in a folder, e.g., ./bert-base-uncased/
Run the program

python -u run_bperp.py \
--cuda \
--data_dir ../Amazon/ \
--index_dir ../Amazon/2/ \
--model_name ./bert-base-uncased/ \
--lr 0.0001 >> bperp.log

Friendly reminders

If you want to do follow-up works on our BPER/BPER-J, please modify the code of BPER+, as it is more efficient.
If you do so, please set the maximum iteration number to a relatively large value, e.g., --epochs 50.

Code dependencies

Python 3.6
PyTorch 1.6

Code reference

Tutorial: Fine tuning BERT for Sentiment Analysis

Citations

@article{TIST22-BPER,
	title={On the Relationship between Explanation and Recommendation: Learning to Rank Explanations for Improved Performance},
	author={Li, Lei and Zhang, Yongfeng and Chen, Li},
	journal={ACM Transactions on Intelligent Systems and Technology (TIST)},
	year={2022}
}
@inproceedings{SIGIR21-EXTRA,
	title={EXTRA: Explanation Ranking Datasets for Explainable Recommendation},
	author={Li, Lei and Zhang, Yongfeng and Chen, Li},
	booktitle={SIGIR},
	year={2021}
}

Recommend Projects