(1) Department of Mathematics, California State University, San Marcos, CA 92096, USA;(2) Informatics and Mathematical Modelling, Technical University of Denmark, Building 321, DK-2800 Lyngby, Denmark
Abstract:
Symmetric VSV decompositions are new rank-revealing decompositions that exploit and preserve symmetry. Truncated VSV solutions are stabilized solutions computed by neglecting blocks in the VSV decomposition with small norm. We compare the truncated VSV solutions with truncated SVD solutions and give perturbation bounds for the VSV solutions. Numerical examples illustrate our results.