Extension of SpotLight
Glaresys is written based on SpotLight. It is an extension of SpotLight. It keeps the main parts of SpotLight except for Sequence Models.
The contributions made here are to add recommender models based on various autoencoders, such as Autoencoder, Denoising Autoencoder, Variational Autoencoder etc., and also to add many evaluation matrics for rating and ranking.
Glaresys is designed for the MSc Project in UofG focusing on Collaborative Autoencoder Recommenders.
In order to lay a solid foundation of for the MSc Project. Basic models are implemented in the system, such as Autoencoder model, Denoising Autoencoder model, Variational Autoencoder model.
'Matrix Factorization' model is extended to deal with item_based prediction and ranking by adding 'item_based_train_test_split' function.
Both 'Autoencoder' and 'Denoising Autoencoder' models can use U_based model and I_based models for rating and ranking .
Besides the fundamental work mentioned above, I also implement 'Hybrid Variational Autoencoder' and 'Conditional Variational Autoencoder' models.
Inspired by these two models, a new model called 'Hybrid Conditional Variational Autoencoder' model is proposed for implicit recommendation.
pip install glaresys
- glaresys/evaluation.py/functions( _get_average_precision, precision_recall_meanofall_score, average_precision_score, map_score, ae_rmse_score, ae_mae_score, ae_precision_recall_score, ae_precision_recall_meanofall_score, ae_average_precision_score, ae_map_score, rr_score, mrr_score, ae_rr_score, ae_mrr_score, ndcg_score, ndcg_meanofall_score, ae_ndcg_score, ae_ndcg_meanofall_score, )
- glaresys/cross_validation.py/functions( item_based_train_test_split, cross_val_score_r, cross_val_score_u, cross_val_score_i, cross_val_score, user_based_general_sparse_split, )
- glaresys/losses/functions( BCE_KLD_loss )
- fix bugs
- Yaxiong Wu (Yashon) from School of Computing Science at University of Glasgow
Release history Release notifications
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
|Filename, size||File type||Python version||Upload date||Hashes|
|Filename, size glaresys-0.1.23-py3-none-any.whl (77.6 kB)||File type Wheel||Python version py3||Upload date||Hashes View hashes|
|Filename, size glaresys-0.1.23.tar.gz (49.2 kB)||File type Source||Python version None||Upload date||Hashes View hashes|
Hashes for glaresys-0.1.23-py3-none-any.whl