My goal is to find a clustering (or an embedding in a vector space) algorithm for a citation network with nodes containing rdf (resource description framework) data (description logic).
Dyadic relational data has a natural representation as an adjacency tensor of third order but I haven't found a way to represent a citation network described above.
I googled for a coupling between a tensor and a matrix and coupling between tensors but google doesn't retrieve any useful result.
I also googled for clustering algorithms working on a coupled tensor-matrix but the best retrieved paper was Coupled Graphs and Tensor Factorization for Recommender Systems and Community Detection, which doesn't solve my problem.
My question is:
Is there a nonnegative matrix factorization algorithm for clustering a citation network?