Maximum likelihood parameter estimation for latent variable models using sequential Monte Carlo

AM Johansen; A Doucet; M Davy

doi:10.1109/ICASSP.2006.1660735

Maximum likelihood parameter estimation for latent variable models using sequential Monte Carlo

AM Johansen, A Doucet, M Davy

School of Mathematics

Research output: Chapter in Book/Report/Conference proceeding › Conference Contribution (Conference Proceeding)

3 Citations (Scopus)

Abstract

We present a sequential Monte Carlo (SMC) method for maximum likelihood (ML) parameter estimation in latent variable models. Standard methods rely on gradient algorithms such as the Expectation-Maximization (EM) algorithm and its Monte Carlo variants. Our approach is different and motivated by similar considerations to simulated annealing (SA); that is we propose to sample from a sequence of artificial distributions whose support concentrates itself on the set of ML estimates. To achieve this we use SMC methods. We conclude by presenting simulation results on a toy problem and a nonlinear non-Gaussian time series model.

Translated title of the contribution	Maximum likelihood parameter estimation for latent variable models using sequential Monte Carlo
Original language	English
Title of host publication	2006 IEEE International Conference on Acoustics Speech and Signal Processing Proceedings
Publisher	Institute of Electrical and Electronics Engineers (IEEE)
Pages	640 - 643
Number of pages	4
ISBN (Print)	1-4244-0469-X
DOIs	https://doi.org/10.1109/ICASSP.2006.1660735
Publication status	Published - May 2006

Publication series

Name
ISSN (Print)	1520-6149

Bibliographical note

ISBN: 142440469X
Publisher: IEEE
Name and Venue of Conference: IEEE International Conference on Acoustics, Speech and Signal Processing, ICASSP 2006, Toulouse, 14-19 May
Conference Organiser: IEEE

Access to Document

10.1109/ICASSP.2006.1660735

http://tinyurl.com/2zcv8q

Cite this

@inproceedings{3cfea59cb1ac4171a356de6cbab3551a,

title = "Maximum likelihood parameter estimation for latent variable models using sequential Monte Carlo",

abstract = "We present a sequential Monte Carlo (SMC) method for maximum likelihood (ML) parameter estimation in latent variable models. Standard methods rely on gradient algorithms such as the Expectation-Maximization (EM) algorithm and its Monte Carlo variants. Our approach is different and motivated by similar considerations to simulated annealing (SA); that is we propose to sample from a sequence of artificial distributions whose support concentrates itself on the set of ML estimates. To achieve this we use SMC methods. We conclude by presenting simulation results on a toy problem and a nonlinear non-Gaussian time series model.",

author = "AM Johansen and A Doucet and M Davy",

note = "ISBN: 142440469X Publisher: IEEE Name and Venue of Conference: IEEE International Conference on Acoustics, Speech and Signal Processing, ICASSP 2006, Toulouse, 14-19 May Conference Organiser: IEEE",

year = "2006",

month = may,

doi = "10.1109/ICASSP.2006.1660735",

language = "English",

isbn = "1-4244-0469-X",

publisher = "Institute of Electrical and Electronics Engineers (IEEE)",

pages = "640 -- 643",

booktitle = "2006 IEEE International Conference on Acoustics Speech and Signal Processing Proceedings",

address = "United States",

}

Maximum likelihood parameter estimation for latent variable models using sequential Monte Carlo. / Johansen, AM; Doucet, A; Davy, M.

2006 IEEE International Conference on Acoustics Speech and Signal Processing Proceedings. Institute of Electrical and Electronics Engineers (IEEE), 2006. p. 640 - 643.

Research output: Chapter in Book/Report/Conference proceeding › Conference Contribution (Conference Proceeding)

TY - GEN

T1 - Maximum likelihood parameter estimation for latent variable models using sequential Monte Carlo

AU - Johansen, AM

AU - Doucet, A

AU - Davy, M

N1 - ISBN: 142440469X Publisher: IEEE Name and Venue of Conference: IEEE International Conference on Acoustics, Speech and Signal Processing, ICASSP 2006, Toulouse, 14-19 May Conference Organiser: IEEE

PY - 2006/5

Y1 - 2006/5

N2 - We present a sequential Monte Carlo (SMC) method for maximum likelihood (ML) parameter estimation in latent variable models. Standard methods rely on gradient algorithms such as the Expectation-Maximization (EM) algorithm and its Monte Carlo variants. Our approach is different and motivated by similar considerations to simulated annealing (SA); that is we propose to sample from a sequence of artificial distributions whose support concentrates itself on the set of ML estimates. To achieve this we use SMC methods. We conclude by presenting simulation results on a toy problem and a nonlinear non-Gaussian time series model.

AB - We present a sequential Monte Carlo (SMC) method for maximum likelihood (ML) parameter estimation in latent variable models. Standard methods rely on gradient algorithms such as the Expectation-Maximization (EM) algorithm and its Monte Carlo variants. Our approach is different and motivated by similar considerations to simulated annealing (SA); that is we propose to sample from a sequence of artificial distributions whose support concentrates itself on the set of ML estimates. To achieve this we use SMC methods. We conclude by presenting simulation results on a toy problem and a nonlinear non-Gaussian time series model.

U2 - 10.1109/ICASSP.2006.1660735

DO - 10.1109/ICASSP.2006.1660735

M3 - Conference Contribution (Conference Proceeding)

SN - 1-4244-0469-X

SP - 640

EP - 643

BT - 2006 IEEE International Conference on Acoustics Speech and Signal Processing Proceedings

PB - Institute of Electrical and Electronics Engineers (IEEE)

ER -

Maximum likelihood parameter estimation for latent variable models using sequential Monte Carlo

Abstract

Publication series

Bibliographical note

Access to Document

Fingerprint

Cite this