Hi all, I am a little confused about how collaborative filtering depends on other users preferences. Does this happen because we have added the term (Ui) at the end of the predicted rating expression (Week 2, Slide 22)? Because if this term is not there, then the user’s rating would entirely depend on his weights for each feature of the movie and various features of the movie, and there would be no role for other users preferences. Is my understanding correct?
Yes, you are correct.