Movie Recommendation - About

Type in a Movie you like and get recommendations! The recommendations are based on unique feature vectors, created using a Cuda implementation of Non-negative matrix factorisation, applied to the Movielens dataset. The title search is based on the edit distance and is quite forgiving; you can include the year to avoid ambiguities (e.g. Ghostbusters 1984).

All movie posters and descriptions are sourced from and are licended under CC BY-NC 4.0.