Movie Recommendation System

In this project we are comparing two approaches for movie recommendation for a new user or existing user based on their age, gender and occupation.

We have implemented all the methods from scratch and tried to improve the existing movie recommender system.

Approach 1:

SVD-Collaborative Filtering (Clustering existing users in the system and then recommending movies)

Approach 2:

k-clique method (Clustering existing users in the system using connected graph and then recommending movies)

Result:

RMSE value improved by 12% by using Approach-2 (K-clique method)