
SVD AND SIGNAL PROCESSING, III
Algorithms, Architectures and Applications
M. Moonen and B. De Moor (Editors)
9 1995 Elsevier Science B.V. All rights reserved.
149
,APPROXIMATING THE PSVD AND QSVD
S. QIAO
Department of Computer Science and Systems
McMaster University
Hamilton, Ontario
Canada LSS ~K1
qiao@maccs, dcss. mcmaster, ca
ABSTRACT. This paper presents an adaptive method for approximating the SVD of the
product
AB
or the quotient
AB -1
of two matrices A and B. Specifically, it computes an
approximation of the complete orthogonal decomposition of
AB
or
AB -1.
The algorithm
is an extension of the