whong92 / recommender

Various implementations of recommender systems
0 stars 0 forks source link

Explore Asymmetric SVD and generalizations #10

Closed whong92 closed 4 years ago

whong92 commented 4 years ago

Describing users as sums of their items is handy - especially for online models that constantly update, user features need not be constantly updated in a training loop.

https://github.com/gpfvic/IRR/blob/master/Factorization%20meets%20the%20neighborhood-%20a%20multifaceted%20collaborative%20filtering%20model.pdf