Optimal multilinear estimation of a random vector under constraints of causality and limited memory

Authors:
P. G. Howlett;A. Torokhti;C. E. M. Pearce
Affiliations:
School of Mathematics and Statistics, University of South Australia, SA 5095, Australia;School of Mathematics and Statistics, University of South Australia, SA 5095, Australia;School of Mathematical Sciences, University of Adelaide, SA 5005, Australia
Venue:
Computational Statistics & Data Analysis
Year:
2007

Citing 11
Cited 2

Introduction to statistical pattern recognition (2nd ed.)

Introduction to statistical pattern recognition (2nd ed.)
Abstract Optimal Linear Filtering

SIAM Journal on Control and Optimization
Efficient Locally Weighted Polynomial Regression Predictions

ICML '97 Proceedings of the Fourteenth International Conference on Machine Learning
A Krylov Subspace Method for Covariance Approximation and Simulation of Random Processes and Fields

Multidimensional Systems and Signal Processing
Estimating the covariance matrix: a new approach

Journal of Multivariate Analysis
Constructing fixed rank optimal estimators with method of best recurrent approximations

Journal of Multivariate Analysis
Empirical Bayesian estimation of normal variances and covariances

Journal of Multivariate Analysis
A well-conditioned estimator for large-dimensional covariance matrices

Journal of Multivariate Analysis
Relative Karhunen-Loeve transform

IEEE Transactions on Signal Processing
An optimal filter of the second order

IEEE Transactions on Signal Processing
Estimating a covariance matrix from incomplete realizations of arandom vector

IEEE Transactions on Signal Processing

Editorial: 2nd Special Issue on Statistical Signal Extraction and Filtering

Computational Statistics & Data Analysis
Data compression under constraints of causality and variable finite memory

Signal Processing

Quantified Score

Hi-index	0.04

Visualization

Abstract

A new technique is provided for random vector estimation from noisy data under the constraints that the estimator is causal and dependent on at most a finite number p of observations. Nonlinear estimators defined by multilinear operators of degree r are employed, the choice of r allowing a trade-off between the accuracy of the optimal filter and the complexity of the calculations. The techniques utilise an exact correspondence of the nonlinear problem to a corresponding linear one. This is then solved by a new procedure, the least squares singular pivot algorithm, whereby the linear problem can be repeated reduced to smaller structurally similar problems. Invertibility of the relevant covariance matrices is not assumed. Numerical experiments with real data are used to illustrate the efficacy of the new algorithm.