Variational Bayesian inference for partially observed stochastic dynamical systems - art. no. 012022

B. Wang; D. M. Titterington

Variational Bayesian inference for partially observed stochastic dynamical systems - art. no. 012022

B. Wang, D. M. Titterington

Mathematics

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

Abstract

In this paper the variational Bayesian approximation for partially observed continuous time stochastic processes is studied. We derive an EM-like algorithm and describe its implementation. The variational Expectation step is explicitly solved using the method of conditional moment generating functions and stochastic partial differential equations. The numerical experiments demonstrate that the variational Bayesian estimate is more robust than the EM algorithm.

Original language	English
Title of host publication	INTERNATIONAL WORKSHOP ON STATISTICAL-MECHANICAL INFORMATICS 2008 (IW-SMI 2008)
Editors	M Hayashi, JI Inoue, Y Kabashima, K Tanaka
Place of Publication	BRISTOL
Publisher	IOP Publishing
Pages	12022-12022
Number of pages	11
ISBN (Print)	*****************
Publication status	Published - 2009
Event	International Workshop on Statistical-Mechanical Informatics - Sendai Duration: 14 Sept 2008 → 17 Sept 2008

Conference

Conference	International Workshop on Statistical-Mechanical Informatics
City	Sendai
Period	14/09/08 → 17/09/08

Keywords

LIKELIHOOD PARAMETER-ESTIMATION
EM ALGORITHM
CONTROL LAWS

Cite this

@inproceedings{5596a1998f7149278ca43cb52a7d24c5,

title = "Variational Bayesian inference for partially observed stochastic dynamical systems - art. no. 012022",

abstract = "In this paper the variational Bayesian approximation for partially observed continuous time stochastic processes is studied. We derive an EM-like algorithm and describe its implementation. The variational Expectation step is explicitly solved using the method of conditional moment generating functions and stochastic partial differential equations. The numerical experiments demonstrate that the variational Bayesian estimate is more robust than the EM algorithm.",

keywords = "LIKELIHOOD PARAMETER-ESTIMATION, EM ALGORITHM, CONTROL LAWS",

author = "B. Wang and Titterington, {D. M.}",

year = "2009",

language = "English",

isbn = "*****************",

pages = "12022--12022",

editor = "M Hayashi and JI Inoue and Y Kabashima and K Tanaka",

booktitle = "INTERNATIONAL WORKSHOP ON STATISTICAL-MECHANICAL INFORMATICS 2008 (IW-SMI 2008)",

publisher = "IOP Publishing",

address = "United Kingdom",

note = "International Workshop on Statistical-Mechanical Informatics ; Conference date: 14-09-2008 Through 17-09-2008",

}

Wang, B & Titterington, DM 2009, Variational Bayesian inference for partially observed stochastic dynamical systems - art. no. 012022. in M Hayashi, JI Inoue, Y Kabashima & K Tanaka (eds), INTERNATIONAL WORKSHOP ON STATISTICAL-MECHANICAL INFORMATICS 2008 (IW-SMI 2008). IOP Publishing, BRISTOL, pp. 12022-12022, International Workshop on Statistical-Mechanical Informatics, Sendai, 14/09/08.

Variational Bayesian inference for partially observed stochastic dynamical systems - art. no. 012022. / Wang, B.; Titterington, D. M.
INTERNATIONAL WORKSHOP ON STATISTICAL-MECHANICAL INFORMATICS 2008 (IW-SMI 2008). ed. / M Hayashi; JI Inoue; Y Kabashima; K Tanaka. BRISTOL: IOP Publishing, 2009. p. 12022-12022.

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

TY - GEN

T1 - Variational Bayesian inference for partially observed stochastic dynamical systems - art. no. 012022

AU - Wang, B.

AU - Titterington, D. M.

PY - 2009

Y1 - 2009

N2 - In this paper the variational Bayesian approximation for partially observed continuous time stochastic processes is studied. We derive an EM-like algorithm and describe its implementation. The variational Expectation step is explicitly solved using the method of conditional moment generating functions and stochastic partial differential equations. The numerical experiments demonstrate that the variational Bayesian estimate is more robust than the EM algorithm.

AB - In this paper the variational Bayesian approximation for partially observed continuous time stochastic processes is studied. We derive an EM-like algorithm and describe its implementation. The variational Expectation step is explicitly solved using the method of conditional moment generating functions and stochastic partial differential equations. The numerical experiments demonstrate that the variational Bayesian estimate is more robust than the EM algorithm.

KW - LIKELIHOOD PARAMETER-ESTIMATION

KW - EM ALGORITHM

KW - CONTROL LAWS

M3 - Conference contribution

SN - *****************

SP - 12022

EP - 12022

BT - INTERNATIONAL WORKSHOP ON STATISTICAL-MECHANICAL INFORMATICS 2008 (IW-SMI 2008)

A2 - Hayashi, M

A2 - Inoue, JI

A2 - Kabashima, Y

A2 - Tanaka, K

PB - IOP Publishing

CY - BRISTOL

T2 - International Workshop on Statistical-Mechanical Informatics

Y2 - 14 September 2008 through 17 September 2008

ER -