Recommender-Systems-for-Implicit-Feedback-datasets
Recommender-Systems-for-Implicit-Feedback-datasets copied to clipboard
could you please list the papers that explain the loss function
The loss function in the part Augmented Alternating Least Squares (Augmented-ALS) is quite puzzle to me, would you mind attaching the links of papers that related to this loss function?
more specifically, I wish to know this part:
why can we add the auG to the latent factor vector? As auG is about the user information while xu is to the User-Item matrix
Thanks in advance!
hi,, this is one way of augmenting the latent factor with a projection of the user/item's features. A similar approach was used in https://arxiv.org/abs/1609.02116 Hope this helps clarify the loss function
Great! Many thanks to you!