In this work, we address the problem of contextual recommendations by exploiting the concept of subspace clustering. Specifically, we pre-partition users that have rated subsets of data items similarly into clusters and ...