Recommendations with IBM

Udacity DataScientist Nanodegree

Andrzej Wodecki, 09.2019

Project overview

The goal of this project is to:

1. analyze user-article interaction in and IBM Watson Community
2. perform rank and user-user based recommendation
3. perform matrix factorization
4. evaluate recommendations on train and test dataset
5. suggest next steps to improve the algorithm.

Implementation

Just open and run commands in Recommendations_with_IBM.ipynb.

Requirements

You will need basic libraries: numpy==1.15.4 pandas==0.22.0 matplotlib pickle

All other accessories (data and test files) are provided in this repo.

Acknowledgments

1. Udacity.com: for a great idea for the project, and a 'starter' pack (useful scripts)
2. IBM.com for it's good datasets on user-article interactions.