/SYMA_project

Primary LanguageJupyter Notebook

SYMA

authors:

  • Moustapha Diop
  • Mathieu Rivier
  • Marc Monteil
  • Theo Perinet
  • Pierre-Louis Landouzi

To run our code, you just need to open the notebooks in jupyter and run each cells. If you just need to execute our code to have the best results, run the "Hybride Filtering - Page Rank x Cosine Similarity.ipynb" notebook.

Folder content:

  • data/ => contain the datas:
    • dressipi_recsys2022.zip => zip file containing the given files
    • get_data.py => python file used to get datas from the previous zip file
  • functions/ => contain the functions:
    • page_rank.py => python file containing all the algorithms and models
  • matrix_saves/ => directory used to store important results that are very long to cumpute
  • Collaborative Filtering - Page Rank.ipynb => Recommendation using the Page Rank algorithm
  • Content-Based Filtering - Bayes.ipynb => Recommendation using a Bayes predictor
  • Content-Based Filtering - Cosine Similarity.ipynb => Recommendation using the Cosine Similarity algorithm
  • Hybride Filtering - Page Rank x Cosine Similarity.ipynb => Recommendation using the Page Rank and Cosine Similarity algorithms