Symmetric Tucker Decomposition — tucker

Computes a Tucker decomposition with a single orthonormal factor matrix shared by every mode via higher-order orthogonal iteration on the symmetric subspace, mirroring the MATLAB Tensor Toolbox tucker_sym.

Usage

tucker_sym(
  X,
  r,
  tol = 1e-06,
  maxiters = 100L,
  init = "nvecs",
  symmetrize = FALSE
)

Arguments

X: A symmetric Tensor (all modes the same size).
r: Target subspace rank.
tol: Convergence tolerance on change in core norm (default 1e-6).
maxiters: Maximum number of iterations (default 100).
init: "nvecs" (default) or an initial n x r matrix with orthonormal columns.
symmetrize: Logical; if TRUE, symmetrize X first.

Value

A TTensor whose factor matrices are all identical.

Examples

set.seed(1)
X <- symmetrize(tensor(array(rnorm(27), dim = c(3, 3, 3))))
T <- tucker_sym(X, r = 2)