Skip to content

This repo contains the implementation of netflix's original recommender system based on matrix factorization which changed the recommender systems forever 10 years back. Delve in to the notebooks to learn the intricacies of the magic box that knows your exact taste.

License

Notifications You must be signed in to change notification settings

Rohan-Thoma/Netflix-Movie-Recommendation

Repository files navigation

Netflix-Movie-Recommendation

netflix-q.jpg

Notebook Guide:

  • EDA_and_preprocessing.ipynb --> This shows the exploratory data analysis of the netflix data
  • Modelling.ipynb --> This shows the implementation of various traditional machine learning models along with the implementation of matrix factorization in the end, including the various variations which are given below with the formula.

Models roadmap which is followed

models.jpg

SVD matrix factorization user movie interactions

Predicted rating

1.png

Optimization problem with user item interactions and regularization ( to avoid overfitting )

2.png

SVD Matrix factorization with implicit feedback from user ( user rated movies )

3.png

Optimization problem with user item interactions and regularization ( to avoid overfitting )

4.png

Final Results

Screenshot-2024-02-12-091910.png

About

This repo contains the implementation of netflix's original recommender system based on matrix factorization which changed the recommender systems forever 10 years back. Delve in to the notebooks to learn the intricacies of the magic box that knows your exact taste.

Topics

Resources

License

Stars

Watchers

Forks