开源项目Recommender System with TF2.0
主要是对阅读过的部分推荐系统、CTR预估论文进行复现,包括传统模型(MF、FM、FFM等)、神经网络模型(WDL、DCN等)以及序列模型(DIN)。
建立原因:
- 理论和实践似乎有很大的间隔,学术界与工业界的差距更是如此;
- 更好的理解论文的核心内容,增强自己的工程能力;
- 很多论文给出的开源代码都是TF1.x,因此想要用更简单的TF2.0进行复现;
项目特点:
- 使用Tensorflow2.0进行复现;
- 每个模型都是相互独立的,不存在依赖关系;
- 模型基本按照论文进行构建,实验尽量使用论文给出的的公共数据集;
- 具有【Wiki】,对于模型、实验数据集有详细的介绍和链接;
- 代码源文件参数、函数命名规范,并且带有标准的注释;
1、通过git命令git clone https://github.com/ZiyaoGeng/Recommender-System-with-TF2.0.git
或者直接下载;
2、需要环境Python3.7,Tensorflow2.0;
3、根据自己数据集的位置,合理更改所需模型文件内train.py
的file
路径;
4、设置超参数
,直接运行即可;
Paper|Model | Published in | Author |
---|---|---|
Matrix Factorization Techniques for Recommender Systems|MF | IEEE Computer Society,2009 | Koren|Yahoo Research |
Factorization Machines|FM | ICDM, 2010 | Steffen Rendle |
Field-aware Factorization Machines for CTR Prediction|FFM | RecSys, 2016 | Yuchin Juan|Criteo Research |
Paper|Model | Published in | Author |
---|---|---|
Wide & Deep Learning for Recommender Systems|WDL | DLRS, 2016 | Google Inc. |
Deep Crossing: Web-Scale Modeling without Manually Crafted Combinatorial Features|Deep Crossing | KDD, 2016 | Microsoft Research |
Product-based Neural Networks for User Response Prediction|PNN | ICDM, 2016 | Shanghai Jiao Tong University |
Deep & Cross Network for Ad Click Predictions|DCN | ADKDD, 2017 | Stanford University|Google Inc. |
Neural Factorization Machines for Sparse Predictive Analytics|NFM | SIGIR, 2017 | Xiangnan He |
Neural network-based Collaborative Filtering|NCF | WWW, 2017 | Xiangnan He |
Attentional Factorization Machines: Learning the Weight of Feature Interactions via Attention Networks|AFM | IJCAI, 2017 | Zhejiang University|National University of Singapore |
DeepFM: A Factorization-Machine based Neural Network for CTR Prediction|DeepFM | IJCAI, 2017 | Harbin Institute of Technology|Noah’s Ark Research Lab, Huawei |
xDeepFM: Combining Explicit and Implicit Feature Interactions for Recommender Systems|xDeepFM | KDD, 2018 | University of Science and Technology of China |
Paper|Model | Published in | Author |
---|---|---|
Deep Interest Network for Click-Through Rate Prediction|DIN | KDD, 2018 | Alibaba Group |
1、对于项目有任何建议或问题,可以在Issue
留言,或者可以添加作者微信zgzjhzgzy
。
2、作者有一个自己的公众号:推荐算法的小齿轮,如果喜欢里面的内容,不妨点个关注。