Title :
Order Selection for Non-Stationary AR Models by Predictive Least-Squares
Author :
Hambaba, Mohamed L.
Author_Institution :
Stevens Institute of Technology, EECS Department, Hoboken, NJ 07030
Abstract :
We present a new criterion for selecting the order of a non-stationary AR model. It is based on the predictive least-squares principles and is implemented in a computational efficient way by predictive lattice filters. We prove the consistency of the criterion.
Keywords :
Adaptive control; Adaptive equalizers; Communication channels; Computational efficiency; Costs; Filters; Lattices; Least squares approximation; Polynomials; Predictive models;
Conference_Titel :
American Control Conference, 1989
Conference_Location :
Pittsburgh, PA, USA
