Closed ncsilvaa closed 2 years ago
I'm not sure if we should keep other recommendation algorithms in this repo (those in mf or recommenders).
Thanks for you contribution @ncsilvaa! Filtering the things you mentioned these are the things we are going to do:
So, Entropy0 is a literature strategy, this is the actual name of the method (see "A survey of active learning in collaborative filtering recommender systems" or "C. Boutilier, R.S. Zemel, B. Marlin, Active collaborative filtering").
About the matrix factorization, they are important to be part of the package. Implemented literature models use them and maybe others want to use them in a new method, modify, etc.
It's time to clean (making all copies as necessary) the value functions.
Think about the possibility of grouping these value functions into subfolders (like recommendation classes - I can help in this division).