Blind deconvolution of covariance matrix inverses for autoregressive processes

DOI10.1016/J.LAA.2020.02.005MaRDI QIDQ2310396zbMATH OpenOpenAlexFDO

Authors Nina Golyandina, A. Zhigljavsky

Publication date 6 April 2020

Published in Linear Algebra and its Applications (Search for Journal in Brave)

Full work available at URL https://arxiv.org/abs/1909.07216, http://orca.cf.ac.uk/129987/1/blind_deconvolution_withAR%28p%29_february.pdf

zbMATH Keywords

autoregressive processcorrelated noisematrix convolutionstructured low-rank approximation

Mathematics Subject Classification ID

Canonical forms, reductions, classification (15A21) Matrix equations and identities (15A24) Applications of operator theory in systems, signals, circuits, and control theory (47N70) Time series, auto-correlation, regression, etc. in statistics (GARCH) (62M10) Signal theory (characterization, reconstruction, filtering, etc.) (94A12)

Abstract: Matrix $mathbfC$ can be blindly deconvoluted if there exist matrices $mathbfA$ and $mathbfB$ such that $mathbfC=mathbfAastmathbfB$, where $ast$ denotes the operation of matrix convolution. We study the problem of matrix deconvolution in the case where matrix $mathbfC$ is proportional to the inverse of the autocovariance matrix of an autoregressive process. We show that the deconvolution of such matrices is important in problems of Hankel structured low-rank approximation (HSLRA). In the cases of autoregressive models of orders one and two, we fully characterize the range of parameters where such deconvolution can be performed and provide construction schemes for performing deconvolutions. We also consider general autoregressive models of order $p$, where we prove that the deconvolution $mathbfC=mathbfAastmathbfB$ does not exist if the matrix $mathbfB$ is diagonal and its size is larger than $p$.

Recommendations

• Publication:3026094
• A simple method for computing the covariance matrix and its inverse of a stationary autoregressive process
• scientific article; zbMATH DE number 704017
• scientific article; zbMATH DE number 447041
• A simplified approach to inverting the autocovariance matrix of a general \(\mathrm{ARMA}(p,q)\) process

Cites work

• A generalized least-square matrix decomposition
• An approximate inverse for the covariance matrix of moving average and autoregressive processes
• Deconvolution of a discrete uniform distribution
• Exact and Approximate Modeling of Linear Systems
• Iterative algorithms for weighted and unweighted finite-rank time-series approximations
• Low-rank approximation. Algorithms, implementation, applications
• On the Inversion of the Sample Covariance Matrix in a Stationary Autoregressive Process
• Optimization challenges in the structured low rank approximation problem
• Signal enhancement-a composite property mapping algorithm
• Singular Spectrum Analysis with R
• Software for weighted structured low-rank approximation
• Stochastic algorithms for solving structured low-rank matrix approximation problems
• Structured low rank approximation

Cited in

(4)

• On the structure and invertibility of the autocorrelation matrix in the unconstrained FDLMS problem
• Discrete uniform and binomial distributions with infinite support
• Search for weights in the problem of finite-rank signal estimation in the presence of random noise
• ESTIMATION AND BLIND DECONVOLUTION OF AUTOREGRESSIVE SYSTEMS WITH NONSTATIONARY BINARY INPUTS

Describes a project that uses

Uses Software

• SLRA

This page was built for publication: Blind deconvolution of covariance matrix inverses for autoregressive processes

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2310396)