Movie-Recommendation (Project 2)

Instructions

Objective:

-Build movie recommender systems.

goals:

- Demographic: Recommend top trending movies ( no user customization ), if no context about the user is known

- Content based (item-item) :Recommend top 10 similar movies, given a recently watched movie. Make sure you engineer a comprehensive set of features to find most similar movies.

- Collaborative filtering: given a data set of user-movie ratings (feel free to change your data set if you need to), predict a movie rating for a given user and movie. ( use SVD or any matrix factorization technique you prefer)

- Serve your model prediction over a restful api (deployed on docker)

Grading Criteria :

EDA ( plots, comments, feature selection, preprocessing ) : 20

Demographic (Top trends) : 10

Content based (item-item similarity) : 40

Collaborative filtering (SVD, NMF) : 20

docker/API : 10

netflix like UI : +10

AUTOML (NCF) : +5