rgib37190 / daisyRec

A developing recommender system in pytorch. Algorithm: KNN, LFM, SLIM, NeuMF, FM, DeepFM, VAE and so on, which aims to fair comparison for recommender system benchmarks

Overview

DaisyRec is a Python toolkit dealing with rating prediction and item ranking issue.

The name DAISY (roughly :) ) stands for multi-Dimension fAirly compArIson for recommender SYstem.

To get all dependencies, run:

pip install -r requirements.txt

Before running, you need first run:

python setup.py build_ext --inplace

to generate .so or .pyd file used for further import.

Make sure you have a CUDA enviroment to accelarate since these deep-learning models could be based on it. We will consistently update this repo.

DaisyRec handled ranking issue mainly and split recommendation problem into point-wise ones and pair-wise ones so that different loss function are constructed such as BPR, Top-1, Hinge and Cross Entropy. All algorithms already implemented are exhibited below:

use main.py to achieve KPI results calculated by certain algorithm above. For example, you can implement this program to implement BPR-MF:

python main.py --problem_type=pair --algo_name=mf --loss_type=BPR --num_ng=2

All experiments code executed in our paper are exhibited in master branch. Please check out to master branch. Code in dev branch is still under developing.

Datasets

You can download experiment data, and put them into the data folder. All data are available in links below:

• MovieLens 100K
• MovieLens 1M
• MovieLens 10M
• MovieLens 20M
• Netflix Prize Data
• Last.fm
• Book Crossing
• Epinions
• CiteULike
• Amazon-Book
• Amazon-Electronic
• Amazon-Cloth
• Amazon-Music
• Yelp Challenge

TODO list

• user-level time-aware fold-out method
• user-level/item-level/user-item-level N-core
• distinguish N-filter and N-core preprocess method
• weight initialization interface
• a more friendly tuner
• make sure Item2vec work as expected

Cite

Here is a Bibtex entry if you ever need to cite DaisyRec in a research paper (please keep us posted, we would love to know if Daisy was helpful to you)

@inproceedings{sun2020are,
  title={Are We Evaluating Rigorously? Benchmarking Recommendation for Reproducible Evaluation and Fair Comparison},
  author={Sun, Zhu and Yu, Di and Fang, Hui and Yang, Jie and Qu, Xinghua and Zhang, Jie and Geng, Cong},
  booktitle={Proceedings of the 14th ACM Conference on Recommender Systems},
  year={2020}
}

Acknowledgements

We refer to the following repositories to improve our code:

• SliM and KNN-CF parts with RecSys2019_DeepLearning_Evaluation
• SVD++ part with Surprise