vendredi 22 juin 2018

SVD Recommendation System, how to test?

I'm working on an Recommendation System using SVD, and I have some matrix R with all user ratings on songs.

R = U D Vt

I understand how to get the training error, but how do I get the testing error. Specifically, how does a user not in the training set get mapped into the low-dimension space?

Aucun commentaire:

Enregistrer un commentaire