> I don't know if there is code, > probably not, but conceptually that is all that it involves.
Once you factorized your interaction matrix, you can use org.apache.mahout.cf.taste.hadoop.als.RecommenderJob to compute recommendations in parallel. Best, Sebastian
