A new approach to particle based smoothed marginal MAP

S. Saha; Pranab K. Mandal; Arunabha Bagchi

A new approach to particle based smoothed marginal MAP

S. Saha, Pranab K. Mandal, Arunabha Bagchi

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › Academic › peer-review

91 Downloads (Pure)

Abstract

We present here a new method of finding the MAP state estimator from the weighted particles representation of marginal smoother distribution. This is in contrast to the usual practice, where the particle with the highest weight is selected as the MAP, although the latter is not necessarily the most probable state estimate. The method developed here uses only particles with corresponding filtering and smoothing weights. We apply this estimator for finding the unknown initial state of a dynamical system and addressing the parameter estimation problem.

Original language	Undefined
Title of host publication	16th European Signal Processing Conference (EUSIPCO 2008)
Place of Publication	Lausanne
Publisher	EURASIP, European Association for Signal, Speech and Image Processing
Pages	1569102598
Number of pages	5
ISBN (Print)	978-2-8399-0450-6
Publication status	Published - Aug 2008
Event	16th European Signal Processing Conference - Lausanne, Switzerland Duration: 25 Aug 2008 → 29 Aug 2008 Conference number: 16th https://www.eurasip.org/Proceedings/Eusipco/Eusipco2008/index.html

Publication series

Name
Publisher	EURASIP, European Association for Signal, Speech and Image Processing

Conference

Conference	16th European Signal Processing Conference
Abbreviated title	EUSIPCO 2008
Country/Territory	Switzerland
City	Lausanne
Period	25/08/08 → 29/08/08
Internet address	https://www.eurasip.org/Proceedings/Eusipco/Eusipco2008/index.html

Keywords

EWI-13360
IR-64949
METIS-252043

Access to Document

http://www.eurasip.org/Proceedings/Eusipco/Eusipco2008/papers/1569102598.pdf

Cite this

@inproceedings{e10e7e723275451e97c16bd4755ac79f,

title = "A new approach to particle based smoothed marginal MAP",

abstract = "We present here a new method of finding the MAP state estimator from the weighted particles representation of marginal smoother distribution. This is in contrast to the usual practice, where the particle with the highest weight is selected as the MAP, although the latter is not necessarily the most probable state estimate. The method developed here uses only particles with corresponding filtering and smoothing weights. We apply this estimator for finding the unknown initial state of a dynamical system and addressing the parameter estimation problem.",

keywords = "EWI-13360, IR-64949, METIS-252043",

author = "S. Saha and Mandal, {Pranab K.} and Arunabha Bagchi",

year = "2008",

month = aug,

language = "Undefined",

isbn = "978-2-8399-0450-6",

publisher = "EURASIP, European Association for Signal, Speech and Image Processing",

pages = "1569102598",

booktitle = "16th European Signal Processing Conference (EUSIPCO 2008)",

note = "16th European Signal Processing Conference ; Conference date: 25-08-2008 Through 29-08-2008",

url = "https://www.eurasip.org/Proceedings/Eusipco/Eusipco2008/index.html",

}

Saha, S, Mandal, PK & Bagchi, A 2008, A new approach to particle based smoothed marginal MAP. in 16th European Signal Processing Conference (EUSIPCO 2008). EURASIP, European Association for Signal, Speech and Image Processing, Lausanne, pp. 1569102598, 16th European Signal Processing Conference, Lausanne, Switzerland, 25/08/08. <http://www.eurasip.org/Proceedings/Eusipco/Eusipco2008/papers/1569102598.pdf>

A new approach to particle based smoothed marginal MAP. / Saha, S.; Mandal, Pranab K.; Bagchi, Arunabha.
16th European Signal Processing Conference (EUSIPCO 2008). Lausanne: EURASIP, European Association for Signal, Speech and Image Processing, 2008. p. 1569102598.

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › Academic › peer-review

TY - GEN

T1 - A new approach to particle based smoothed marginal MAP

AU - Saha, S.

AU - Mandal, Pranab K.

AU - Bagchi, Arunabha

N1 - Conference code: 16th

PY - 2008/8

Y1 - 2008/8

N2 - We present here a new method of finding the MAP state estimator from the weighted particles representation of marginal smoother distribution. This is in contrast to the usual practice, where the particle with the highest weight is selected as the MAP, although the latter is not necessarily the most probable state estimate. The method developed here uses only particles with corresponding filtering and smoothing weights. We apply this estimator for finding the unknown initial state of a dynamical system and addressing the parameter estimation problem.

AB - We present here a new method of finding the MAP state estimator from the weighted particles representation of marginal smoother distribution. This is in contrast to the usual practice, where the particle with the highest weight is selected as the MAP, although the latter is not necessarily the most probable state estimate. The method developed here uses only particles with corresponding filtering and smoothing weights. We apply this estimator for finding the unknown initial state of a dynamical system and addressing the parameter estimation problem.

KW - EWI-13360

KW - IR-64949

KW - METIS-252043

M3 - Conference contribution

SN - 978-2-8399-0450-6

SP - 1569102598

BT - 16th European Signal Processing Conference (EUSIPCO 2008)

PB - EURASIP, European Association for Signal, Speech and Image Processing

CY - Lausanne

T2 - 16th European Signal Processing Conference

Y2 - 25 August 2008 through 29 August 2008

ER -