A globally convergent matricial algorithm for multivariate spectral estimation
Optimization and Control
2008-09-30 v1
Abstract
In this paper, we first describe a matricial Newton-type algorithm designed to solve the multivariable spectrum approximation problem. We then prove its global convergence. Finally, we apply this approximation procedure to multivariate spectral estimation, and test its effectiveness through simulation. Simulation shows that, in the case of short observation records, this method may provide a valid alternative to standard multivariable identification techniques such as MATLAB's PEM and MATLAB's N4SID.
Cite
@article{arxiv.0809.5024,
title = {A globally convergent matricial algorithm for multivariate spectral estimation},
author = {Federico Ramponi and Augusto Ferrante and Michele Pavon},
journal= {arXiv preprint arXiv:0809.5024},
year = {2008}
}