Metadata-Version: 2.1
Name: mangaki-zero
Version: 1.0
Summary: Mangaki's recommandation algorithms
Home-page: http://research.mangaki.fr
License: GPL-3.0
Keywords: machine learning,recommandation,mangaki,recommender systems,algorithms
Author: Jill-Jênn Vie
Author-email: vie@jill-jenn.net
Requires-Python: >=3.7,<4.0
Classifier: License :: OSI Approved :: GNU General Public License v3 (GPLv3)
Classifier: Programming Language :: Python :: 3
Classifier: Programming Language :: Python :: 3.7
Classifier: Programming Language :: Python :: 3.8
Provides-Extra: deep
Provides-Extra: external
Provides-Extra: fm
Requires-Dist: fastFM; extra == "fm"
Requires-Dist: numpy (>=1.18,<2.0)
Requires-Dist: pandas (>=0.25.2,<0.26.0)
Requires-Dist: pywFM; extra == "fm"
Requires-Dist: scikit-learn; extra == "external"
Requires-Dist: scipy (>=1.4.1,<2.0.0)
Requires-Dist: surprise; extra == "external"
Requires-Dist: tensorflow; extra == "deep"
Project-URL: Documentation, https//github.com/mangaki/zero
Description-Content-Type: text/markdown

# Zero

[![Mangaki Zero's CI status](https://github.com/mangaki/zero/workflows/CI/badge.svg)](https://mangaki/zero/actions)
[![Mangaki Zero's code coverage](https://codecov.io/gh/mangaki/zero/branch/master/graph/badge.svg)](https://codecov.io/gh/mangaki/zero)



Mangaki's recommendation algorithms.

It is tested on Python 3.6, 3.7 and 3.8 over OpenBLAS LP64 & MKL.

## Usage

Most models have the following routines:

    from zero.als import MangakiALS
    model = MangakiALS(nb_components=10)
    model.fit(X, y)
    model.predict(X)

There are a couple of other methods that can be used for online fit, say `model.predict_single_user(work_ids, user_parameters)`.

To run k-fold cross-validation, do:

    python compare.py <path/to/dataset>

## Results

### Mangaki data

![Comparing on Mangaki](results/mangaki.png)

### Movielens data

![Comparing on Movielens](results/movielens.png)

Feel free to use. Under GPLv3 license.

