1 paper
Xu Han, Laurent Albera, Amar Kachenoura +2
In order to compute the best low-rank tensor approximation using the Multilinear Tensor Decomposition (MTD) model, it is essential to estimate the rank of the underlying multilinea…