CF-KAN is a recommendation method (collaborative filtering) based on the Kolmogorov-Arnold Network (KAN) approach. This research leverages and explores the power of KAN for CF-based recommendation.
** Our paper (CF-KAN) is now available at arXiv: https://www.arxiv.org/abs/2409.05878**
python main.py
We comprehensively compare CF-KAN with the following categories of recommendation models: Matrix factorization-based methods: MF-BPR, NeuMF, and DMF; Autoencoder-based methods: CDAE, Multi-DAE, and RecVAE; GCN-based methods: SpectralCF, NGCF ,LightGCN, SGL, and NCL; Generative model-based methods: CFGAN, RecVAE, and DiffRec.
| Method | ML-1M | Yelp | Anime | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| R@10 | R@20 | N@10 | N@20 | R@10 | R@20 | N@10 | N@20 | R@10 | R@20 | N@10 | N@20 | |
| MF-BPR | 0.0876 | 0.1503 | 0.0749 | 0.0966 | 0.0341 | 0.0560 | 0.0210 | 0.0341 | 0.1521 | 0.2449 | 0.2925 | 0.3153 |
| NeuMF | 0.0845 | 0.1465 | 0.0759 | 0.0965 | 0.0378 | 0.0637 | 0.0230 | 0.0308 | 0.1531 | 0.2442 | 0.3277 | 0.3259 |
| DMF | 0.0799 | 0.1368 | 0.0731 | 0.0921 | 0.0342 | 0.0588 | 0.0208 | 0.0282 | 0.1386 | 0.2161 | 0.3277 | 0.3122 |
| CDAE | 0.0991 | 0.1705 | 0.0829 | 0.1078 | 0.0444 | 0.0703 | 0.0280 | 0.0360 | 0.2031 | 0.2845 | 0.4652 | 0.4301 |
| MultiDAE | 0.0975 | 0.1707 | 0.0820 | 0.1046 | 0.0531 | 0.0876 | 0.0316 | 0.0421 | 0.2022 | 0.2882 | 0.4577 | 0.4125 |
| RecVAE | 0.0835 | 0.1422 | 0.0769 | 0.0963 | 0.0493 | 0.0824 | 0.0303 | 0.0403 | 0.2137 | 0.3068 | 0.4105 | 0.4068 |
| SpectralCF | 0.0751 | 0.1291 | 0.0740 | 0.0909 | 0.0368 | 0.0572 | 0.0201 | 0.0298 | 0.1633 | 0.2564 | 0.3102 | 0.3236 |
| NGCF | 0.0864 | 0.1484 | 0.0805 | 0.1038 | 0.0417 | 0.0708 | 0.0255 | 0.0346 | 0.2071 | 0.3043 | 0.3937 | 0.3827 |
| LightGCN | 0.0824 | 0.1419 | 0.0793 | 0.1006 | 0.0505 | 0.0858 | 0.0312 | 0.0417 | 0.2071 | 0.3043 | 0.3937 | 0.3827 |
| SGL | 0.0885 | 0.1575 | 0.0802 | 0.1029 | 0.0564 | 0.0944 | 0.0346 | 0.0462 | 0.1994 | 0.2918 | 0.3748 | 0.3652 |
| NCL | 0.0878 | 0.1471 | 0.0819 | 0.1011 | 0.0535 | 0.0906 | 0.0326 | 0.0438 | 0.2063 | 0.3047 | 0.3919 | 0.3819 |
| CFGAN | 0.0684 | 0.1181 | 0.0663 | 0.0828 | 0.0206 | 0.0347 | 0.0129 | 0.0172 | 0.1946 | 0.2889 | 0.4601 | 0.4289 |
| DiffRec | 0.1021 | 0.1763 | 0.0877 | 0.1131 | 0.0554 | 0.0914 | 0.0345 | 0.0452 | 0.2104 | 0.3012 | 0.5047 | 0.4649 |
| CF-KAN | 0.1065 | 0.1831 | 0.0894 | 0.1152 | 0.0590 | 0.0961 | 0.0359 | 0.0471 | 0.2287 | 0.3261 | 0.5256 | 0.4875 |
If you found this project is helpful for your research, 🌟STAR🌟 this repository
If our work was helpful for your project, cite our work :)
@article{park2024cf,
title={CF-KAN: Kolmogorov-Arnold Network-based Collaborative Filtering to Mitigate Catastrophic Forgetting in Recommender Systems},
author={Park, Jin-Duk and Kim, Kyung-Min and Shin, Won-Yong},
journal={arXiv preprint arXiv:2409.05878},
year={2024} }
For the implementation of KANs, I've referred to the following repos: EfficentKAN, Visualization API