Official implementation of the paper Deep Unified Representation for Heterogeneous Recommendation.
Accepted by the ACM Web Conference 2022 (WWW '22)
中文版算法介绍
In this paper, we use the Douban dataset stored in data.tar.xz.
Please uncompress it (tar -xf data.tar.xz) and put it in the working directory.
Before runing the code, please make sure that you have installed the dependency. You can install them with
pip install -r requirements.txt
Our code is tested on python 3.7.
The next step is to prepare the configuration file. We provides the configurations of our proposed model (DURation) and baselines in configs fold as examples. To reproduce the results report in our paper, one just need change the path to your local path in the configuration.
Then, you can run the program with simple one-line code. Take the DURation model as a example, there is a duration.json file in configs.
python train_hete.py duration
It is worth note that the configuration file must be put in configs. To test the homogeneous models, just replace train_hete.py with train_homo.py. The program will output the results on screen while save the log to a certain path.
Currently, we support the following models:
- DeepMF(2017): Deep Matrix Factorization Models for Recommender Systems
- FISM(2013): Fism: factored item similarity models for top-n recommender systems.
- NAIS(2018): Nais: Neural attentive item similarity model for recommendation.
- DeepFM(2017): DeepFM: a factorization-machine based neural network for CTR prediction
- xDeepFM(2018): xdeepfm: Combining explicit and implicit feature interactions for recommender systems
- AFM(2017): Attentional factorization machines: Learning the weight of feature interactions via attention networks
- DSSM(2013): Learning deep structured semantic models for web search using clickthrough data
- Wide & Deep(2016): Wide & deep learning for recommender systems
- autoInt(2019): Autoint: Automatic feature interaction learning via selfattentive neural networks
- CCCFNet(2012): Cross-domain collaboration recommendation
- DDTCDR(2020): DDTCDR: Deep dual transfer cross domain recommendation
@inproceedings{lu2022deep,
title={Deep Unified Representation for Heterogeneous Recommendation},
author={Lu, Chengqiang and Yin, Mingyang and Shen, Shuheng and Ji, Luo and Liu, Qi and Yang, Hongxia},
booktitle={Proceedings of the ACM Web Conference 2022},
pages={2141--2152},
year={2022}
}