Hello,
i'm trying to make a recommendation system that recommends items from an
input list of items (and a referent dataset, obviously).
To this end, i'm trying use SDVRecommender and
PlusAnonymousConcurrentUserDataModel. I've already set up the
recommendation system for known users with SVD, and it works.
However, if i try to use PlusAnonymousConcurrentUserDataModel, i
constantly get a NoSuchUserException: -9223372036854775808
I've managed to use a recommender based on Pearson correlation, but i 'd
prefer to use SVD.
So, i wonder if it is just possible to use anonymous users and SVD
recommender in Mahout.
Thanks for your answers
Florian