rater is a comparative framework for multimodal recommender systems. It was developed to facilitate the designing, comparing, and sharing of recommendation models.
- easy to use, rebuild and compare
- SOTA model
- classical model and deep model
- model has great influence in the industry
- model hsa been successfully applied by Google, Alibaba, Baidu and other well-known companies
- engineering oriented, not just experimental data validation
- ml-1m: http://files.grouplens.org/datasets/movielens/ml-1m.zip
- delicious-2k: http://files.grouplens.org/datasets/hetrec2011/hetrec2011-delicious-2k.zip
- lastfm-dataset-360K: http://mtg.upf.edu/static/datasets/last.fm/lastfm-dataset-360K.tar.gz
- slashdot: http://snap.stanford.edu/data/soc-Slashdot0902.txt.gz
- epinions: http://snap.stanford.edu/data/soc-Epinions1.txt.gz
- ml-100k: http://files.grouplens.org/datasets/movielens/ml-100k.zip
- Criteo(dac full): https://s3-eu-west-1.amazonaws.com/kaggle-display-advertising-challenge-dataset/dac.tar.gz
- Criteo(dac sample): http://labs.criteo.com/wp-content/uploads/2015/04/dac_sample.tar.gz
pip3 install rater
or
git clone https://github.com/shibing624/rater.git
cd rater
python3 setup.py install
Load the built-in MovieLens 1M dataset (will be downloaded if not cached):
Output:
MAE | RMSE | AUC | NDCG@10 | Recall@10 | Train (s) | Test (s) | |
---|---|---|---|---|---|---|---|
[MF] | 0.7430 | 0.8998 | 0.7445 | 0.0479 | 0.0352 | 0.13 | 1.57 |
For more details, please take a look at our examples.
The models supported are listed below. Why don't you join us to lengthen the list?
model/keywords | paper |
---|---|
GRU4Rec | Session-based Recommendations with Recurrent Neural Networks |
Caser | Personalized Top-N Sequential Recommendation via Convolutional Sequence Embedding |
DIN: Deep Interest Network | [KDD 2018]Deep Interest Network for Click-Through Rate Prediction |
Self-Attention | Next Item Recommendation with Self-Attention |
Hierarchical Attention | Sequential Recommender System based on Hierarchical Attention Networks |
DIEN: Deep Interest Evolution Network | [AAAI 2019]Deep Interest Evolution Network for Click-Through Rate Prediction |
DISN: Deep Session Interest Network | [IJCAI 2019]Deep Session Interest Network for Click-Through Rate Prediction |
model | paper |
---|---|
node2vec | node2vec: Scalable Feature Learning for Networks |
item2vec | ITEM2VEC: Neural item embedding for collaborative filtering |
Airbnb embedding | Real-time Personalization using Embeddings for Search Ranking at Airbnb |
EGES: Enhanced Graph Embedding with Side information | Billion-scale Commodity Embedding for E-commerce Recommendation in Alibaba |
refer: https://zhuanlan.zhihu.com/p/63186101
Your contributions at any level of the library are welcome. If you intend to contribute, please:
- Fork the rater repository to your own account.
- Make changes and create pull requests.
You can also post bug reports and feature requests in GitHub issues.
- [Multilayer Perceptron Based Recommendation]
- [Autoencoder Based Recommendation]
- [CNN Based Recommendation]
- [RNN Based Recommendation]
- [Restricted Boltzmann Machine Based Recommendation]
- [Neural Attention Based Recommendation]
- [Neural AutoRegressive Based Recommendation]
- [Deep Reinforcement Learning for Recommendation]
- [GAN Based Recommendation]
- [Deep Hybrid Models for Recommendation]
- maciejkula/spotlight
- shenweichen/DeepCTR
- Magic-Bubble/RecommendSystemPractice
- nzc/dnn_ctr