playground is an interactive journal of all the work I have done as part of GovTech's Girls in Tech mentorship programme.
My goal for the programme was to get a feel of some basic machine learning algorithms through developing a simple movie recommendation system.
- Unique genres
- Number of ratings per genre
- Average rating for each genre
Both user-based collaborative filtering and item-based collaborative filtering were explored. The classic k-nearest neighbours algorithm was used to train the models.
User-based collaborative filtering:
- Use ratings of most common movies across users to determine user similarity.
- Use users' average ratings of each genre to determine user similarity.
Item-based collaborative filtering:
- Use user ratings to determine movie similarity.
Tech stack: Flask + React