Computation of the low multilinear rank approximations of clusterd tensors

In this paper we proposed an procedure for obtaining the low multilinear rank approximation of large sparse tensors that comes from massive graphs. This procedure involvs the clustering of tensors into subtensors and then obtain the low multilinear rank approximation of each small subtensors separately and combine them to compute a low multilinear rank approximation of entire tensor. the procedure is efficient both in terms of computational speed and memory usage.