Implementation of 5 methods of recommander system on Netflix data.
- Baseline Estimator ( baseline_estimator.py )
- Correlation Based Neighbourhood Model ( correlation_based_neighbourhood_model.py )
- Correlation Based Implicit Neighbourhood Model ( correlation_based_implicit_neighbourhood_model.py )
- SVD++ ( svd_more_more.py )
- Integrated Model ( integrated_model.py )
Models are described in [Koren, 2008] Factorization Meets the Neighborhood: a Multifaceted Collaborative Filtering Model.
The default dataset path is "../Datasets" (relative path from the root of this repository), it can be modified in utils.py.
In this folder it must be a folder named "download" with the content of the archive that can be download here: https://archive.org/download/nf_prize_dataset.tar
The folder is not included in this repository due the size of the files.
Then you can run rating_compiler.py at first to create the different matrices from the files. Then you can run each algorithms.