hadryan/movie_recomendation_system

"Movie Runner" - Collaborative Filtering Based Movie Recommendation System

“Movie Runner”

Collaborative Filtering Based Movie Recommendation System

20/11/2020

Movie Runner is a model that recommends 10 movies to be liked by each person, based on their rating for movies. Netflix Prize Data was used as a dataset. At the same time, the movies we watched and rated were added to the dataset. Collaborative Filtering algorithm is used with the help of SVD (Singular Value Decomposition) method. Similar users and similar movies are detected by the algorithm by assigning a vector to all movies and users. Then, movies liked by users with the same movie taste are recommended to the person. According to the answer to the question "How many points would he/she give it if he/she watched" the movies that each user has not watched before, it finds the points and the highest 10 scores, and recommends these movies to the user. A website was created for the model with the help of Python's Flask framework. When the name of a user in the dataset is entered on this website, 10 movies will be recommended to that user.