A repository containing a recommendation engine for recommending movies to users based on the popular "FunkSVD" matrix factorization technique (made famous by Simon Funk in the Netflix prize competition) as well as content and knowledge-based recommendations for users/movies that are not contained in the FunkSVD training data (e.g. because they are new users and/or movies)
The recommencation engine is implemented as the Recommender
class which is contained in the recommender.py
file. A more instructive introduction on how the recommendation engine works can be found in the funksvd_rec.ipynb
jupyter notebook, where recommendations are made for user/movie data provided by Udacity through the data scientist nanodegree course.
A default modern python data analysis stack is necessary to run this project. In particular, users will need the following packages:
- numpy
- pandas
- matplotlib
- seaborn