Testing composite hypotheses applied to AR-model order estimation; the Akaike-criterion revised

R. Moddemeijer

University of Groningen, Department of Computing Science,
P.O. Box 800, NL-9700 AV Groningen, The Netherlands,
phone: +31.50.363 3940 - fax: +31.50.363 38005 - e-mail: rudy@cs.rug.nl


Akaike's criterion is used to test composite hypotheses; for example to determine the order of AR models. A modification is presented to test composite hypotheses given an upper-bound on the error of the first kind (Neyman-Pearson). The presented theory is applied to AR order estimation and verified by simulations. The experimental results are so good that we consider the AR order estimation problem as solved.

Full paper



Nineteenth Symposium on Information Theory in the Benelux, May 28-29, 1998, Veldhoven, The Netherlands, pp 149-156, eds. P.H.N. de With et. al., Werkgemeenschap Informatie- en Communicatietheorie, Enschede, and IEEE Benelux Chapter on Information Theory, ISBN 90-71048-13-6, BibTeX

Related work

other publications