Dimensionality reduction in higher-order signal processing and rank-\((R_1,R_2,\ldots,R_N)\) reduction in multilinear algebra (Q1888335)

The authors review a multilinear generalization of the singular value decomposition and the best rank-\((R_1,R_2,\dots,R_N)\) approximation of higher-order tensors. Special attention is paid to their use in the estimation of the column space, row space, etc., of a tensor contaminated by noise. It is shown that they are important tools for dimensionality reduction in higher-order signal processing. The authors discuss their applications in independent component analysis, simultaneous matrix diagonalization and subspace variants of algorithms based on higher-order statistics.