• like
• ratings

• clicks
• Watch time

(Recommended popular movies, this will not personalize. Just recommend the popular items)

(recommended items to similar users behavior)

• memory based (similar users behavior used matrix factorization)

  • user based (find similar user like this)
  • product based (find similar products like this)

• Model based (this is using machine learning model to predict/get hidden patterns underlaying on data)

(recommend the video user previously liked , recomend the similar kind of videos)

• Liked videos
• Commented

Dataset , I am using ml-latest-small.zip small one (1 mb)

Ps : In the project I will use movie-dataset , so the context is movie recommendation system. In above I just go through the basic idea as whole