A new order selection method for auto-regressive processes

An approximate formula for the residual variance of autoregressive modeling with the least-squares-forward (LSF) method is derived. By using this formula the statistical behavior of the residual variance is determined approximately. Based on this formula, a new sequential testing algorithm for order...

Full description

Saved in:
Bibliographic Details
Main Authors: Karimi, M., Bastani, M.H.
Format: Conference Proceeding
Language:English
Subjects:
Artificial intelligence
Computer simulation
Geophysical signal processing
Geophysics
Marine
Random variables
Sequential analysis
Signal processing
Signal processing algorithms
Sonar
Speech processing
Online Access:Request full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:An approximate formula for the residual variance of autoregressive modeling with the least-squares-forward (LSF) method is derived. By using this formula the statistical behavior of the residual variance is determined approximately. Based on this formula, a new sequential testing algorithm for order selection of auto-regressive processes is proposed. The performance of the algorithm is demonstrated by computer simulations.
DOI:10.1109/OCEANS.1998.726302