Journal Article

A minimum-eigenvalue ratio test of the product-moment matrix for time-series model-order estimates

Huixin Chen, P. P. G. Dyke and Hong Zhao

in IMA Journal of Mathematical Control and Information

Published on behalf of Institute of Mathematics and its Applications

Volume 17, issue 1, pages 43-55
Published in print March 2000 | ISSN: 0265-0754
Published online March 2000 | e-ISSN: 1471-6887 | DOI: http://dx.doi.org/10.1093/imamci/17.1.43
A minimum-eigenvalue ratio test of the product-moment matrix for time-series model-order estimates

Show Summary Details

Preview

The near singularity of the product-moment matrix of observed input-output data provides a quick way of checking the order of a linear system. However the technique used so far becomes insensitive when significant amounts of extraneous noise is present. A minimum-eigenvalue ratio test (MERT) is introduced based on the product-moment matrix and some prior knowledge about the system noise. It is used to overcome the problems both of insensitivity arising from system noise and of high computational expense. Both the theoretical analysis and simulation results show that it is a simple and efficient way to estimate the order of linear and some nonlinear time-series models prior to parameter identification.

Keywords: order estimate; minimum-eigenvalue ratio test (MERT); modelling; productmoment matrix; linear system; bilinear system

Journal Article.  0 words. 

Subjects: Mathematics

Full text: subscription required

How to subscribe Recommend to my Librarian

Users without a subscription are not able to see the full content. Please, subscribe or login to access all content.