An efficient bayes solution to AR signal modelling for short sequences

Douglas E. Johnston; Petar M. Djurić

doi:10.1109/ICASSP.1994.389809

An efficient bayes solution to AR signal modelling for short sequences

Douglas E. Johnston
, Petar M. Djurić

Electrical Engineering

Stony Brook University

Research output: Contribution to journal › Conference article › peer-review

Abstract

A Bayesian approach to autoregressive (AR) signal modelling is proposed. In contrast to previous research, the exact posterior density of the model parameters is utilized and minimum mean square estimates (MMSE) are evaluated. To compute the estimates, a numerically efficient procedure is presented which can be viewed as an alternative to multidimensional optimization. Our approach can be used to investigate many signal characteristics such as the signal's spectrum, marginal densities for prediction or even model selection. Simulation results confirm our expectations and illustrate the improvement over the classic, maximum conditional likelihood (MCL) approach to AR signal modelling.

Original language	English
Article number	389809
Pages (from-to)	IV345-IV348
Journal	ICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings
Volume	4
DOIs	https://doi.org/10.1109/ICASSP.1994.389809
State	Published - 1994
Event	Proceedings of the 1994 IEEE International Conference on Acoustics, Speech and Signal Processing. Part 2 (of 6) - Adelaide, Aust Duration: Apr 19 1994 → Apr 22 1994

Access to Document

10.1109/ICASSP.1994.389809

Cite this

@article{ca076d9e3959487ca469b9936a56d211,

title = "An efficient bayes solution to AR signal modelling for short sequences",

abstract = "A Bayesian approach to autoregressive (AR) signal modelling is proposed. In contrast to previous research, the exact posterior density of the model parameters is utilized and minimum mean square estimates (MMSE) are evaluated. To compute the estimates, a numerically efficient procedure is presented which can be viewed as an alternative to multidimensional optimization. Our approach can be used to investigate many signal characteristics such as the signal's spectrum, marginal densities for prediction or even model selection. Simulation results confirm our expectations and illustrate the improvement over the classic, maximum conditional likelihood (MCL) approach to AR signal modelling.",

author = "Johnston, \{Douglas E.\} and Djuri{\'c}, \{Petar M.\}",

note = "Publisher Copyright: {\textcopyright} 1994 IEEE.; Proceedings of the 1994 IEEE International Conference on Acoustics, Speech and Signal Processing. Part 2 (of 6) ; Conference date: 19-04-1994 Through 22-04-1994",

year = "1994",

doi = "10.1109/ICASSP.1994.389809",

language = "English",

volume = "4",

pages = "IV345--IV348",

journal = "ICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings",

issn = "0736-7791",

}

TY - JOUR

T1 - An efficient bayes solution to AR signal modelling for short sequences

AU - Johnston, Douglas E.

AU - Djurić, Petar M.

PY - 1994

Y1 - 1994

N2 - A Bayesian approach to autoregressive (AR) signal modelling is proposed. In contrast to previous research, the exact posterior density of the model parameters is utilized and minimum mean square estimates (MMSE) are evaluated. To compute the estimates, a numerically efficient procedure is presented which can be viewed as an alternative to multidimensional optimization. Our approach can be used to investigate many signal characteristics such as the signal's spectrum, marginal densities for prediction or even model selection. Simulation results confirm our expectations and illustrate the improvement over the classic, maximum conditional likelihood (MCL) approach to AR signal modelling.

AB - A Bayesian approach to autoregressive (AR) signal modelling is proposed. In contrast to previous research, the exact posterior density of the model parameters is utilized and minimum mean square estimates (MMSE) are evaluated. To compute the estimates, a numerically efficient procedure is presented which can be viewed as an alternative to multidimensional optimization. Our approach can be used to investigate many signal characteristics such as the signal's spectrum, marginal densities for prediction or even model selection. Simulation results confirm our expectations and illustrate the improvement over the classic, maximum conditional likelihood (MCL) approach to AR signal modelling.

UR - https://www.scopus.com/pages/publications/85079112569

U2 - 10.1109/ICASSP.1994.389809

DO - 10.1109/ICASSP.1994.389809

M3 - Conference article

AN - SCOPUS:85079112569

SN - 0736-7791

VL - 4

SP - IV345-IV348

JO - ICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings

JF - ICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings

M1 - 389809

T2 - Proceedings of the 1994 IEEE International Conference on Acoustics, Speech and Signal Processing. Part 2 (of 6)

Y2 - 19 April 1994 through 22 April 1994

ER -

An efficient bayes solution to AR signal modelling for short sequences

Abstract

Access to Document

Other files and links

Fingerprint

Cite this