Skip to content

Latest commit

 

History

History
39 lines (29 loc) · 1.73 KB

README.md

File metadata and controls

39 lines (29 loc) · 1.73 KB
Raw

Recommendations with IBM

This project analyzes a real data set including articles and user interactions from the IBM Watson Studio platform and makes recommendations for articles to users in a jupyter notebook environment with following content:

  1. Exploratory Data Analysis
  2. Rank Based Recommendations
  3. User-User Based Collaborative Filtering
  4. Matrix Factorization
  5. Extras & Conclusion

This recommendation system could be used for example to recommend articles on the dashboard of the IBM Watson Studio platform.

Configuration

  • Jupyter notebook environment based on python 3.8 with libraries pandas, numpy, matplotlib.pyplot, pickle and sklearn

Installation on local computer

Download zip folder on local computer, extract files and open "Recommendations_with_IBM.ipynb" with jupyter notebook.

File Manifest

Below you can find the file content of this project:



      - data
      |- articles_community.csv (csv file with data to process)
      |- user-item-interactions.csv (csv file with data to process)

      - project_test.py (python file with test functions)
      - README.md (markdown file with instructions)
      - Recommendations_with_IBM.ipynb (jupyter notebook file with recommendation functions)
      - Recommendations_with_IBM.html (html file with recommendation functions)
      - top_5.p (p file with data of top 5 recommendations for testing of results)
      - top_10.p (p file with data of top 10 recommendations for testing of results)
      - top_20.p (p file with data of top 20 recommendations for testing of results)

Contact

Author: Eugen Iftimoaie

For questions feel free to contact me on my e-mail adress: eugen.iftimoaie@gmx.de