Traditionally, singular value decomposition (SVD) can be used to learn latent feature of user and items according to user-item rating matrix.

Recently, researchers use embedding layers as the input layer of a deep neural network to learn latent representation of entities, like users or items.

What is the difference between these two? What is the possible advantages of embedding method over SVD?

  • $\begingroup$ Well one important advantage of neural networks over SVD (and many other linear models) is, it is able to learn non-linear features. $\endgroup$ – dontloo Nov 5 '16 at 8:25
  • $\begingroup$ Perhaps this paper could help a bit. $\endgroup$ – turdus-merula Nov 11 '16 at 13:57

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.