PERTURBATION THEORY FOR THE ANALYSIS OF SVD-BASED ALGORITHMS.
Document Type
Conference Proceeding
Date of Original Version
1-1-1987
Abstract
The problem of statistically analyzing the performance of signal processing algorithms which use the singular value decomposition (SVD) is addressed. Such decomposition, which is widely used in system identification and parameter estimation, is a nonlinear operation. Consequently, when applied to random data, statistical results are extremely difficult to obtain. The first-order Taylor series expansion is generally used in computing the statistics, but the derivative term makes the statistical analysis very difficult. A power-like method which results in a simple expression is proposed. The singular vector perturbation using both approaches for the case of low-rank approximation is examined.
Publication Title, e.g., Journal
ICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings
Citation/Publisher Attribution
Vaccaro, Richard J., and Alex C. Kot. "PERTURBATION THEORY FOR THE ANALYSIS OF SVD-BASED ALGORITHMS.." ICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings (1987): 1613-1616. https://digitalcommons.uri.edu/ele_facpubs/1200