Title :
A novel algorithm for computing rank reduced matrix approximations
Author :
Manton, Jonathan H. ; Hua, Yingbo
Author_Institution :
Dept. of Electr. & Electron. Eng., Melbourne Univ., Parkville, Vic., Australia
Abstract :
The problem of approximating a matrix by one of lower rank occurs in a number of signal processing problems, including reduced rank Wiener filtering and reduced rank maximum likelihood estimation. This paper derives a novel algorithm for computing the optimal rank-reduced approximation of a given matrix under an arbitrary weighted norm. Simulations demonstrate the advantages of the algorithm over the traditional alternating projections algorithm
Keywords :
Wiener filters; approximation theory; filtering theory; matrix algebra; maximum likelihood estimation; signal processing; Wiener filtering; arbitrary weighted norm; maximum likelihood estimation; rank-reduced matrix approximations; signal processing; Algorithm design and analysis; Approximation algorithms; Computational efficiency; Computational modeling; Cost function; Geometry; Maximum likelihood estimation; Projection algorithms; Signal processing algorithms; Wiener filter;
Conference_Titel :
Wireless Communications, 2001. (SPAWC '01). 2001 IEEE Third Workshop on Signal Processing Advances in
Conference_Location :
Taiwan
Print_ISBN :
0-7803-6720-0
DOI :
10.1109/SPAWC.2001.923922
