Can Latent Features Be Interpreted as Users in Matrix Factorization-Based Recommender Systems? | IEEE Conference Publication | IEEE Xplore