Fast Bayes and the dynamic junction forest

J. Q. Smith; K. N. Papamichail

Fast Bayes and the dynamic junction forest

Research output: Contribution to journal › Article › peer-review

Abstract

It has been shown that junction tree algorithms can provide a quick and efficient method for propagating probabilities in complex multivariate problems when they can be described by a fixed conditional independence structure. In this paper we formalise and illustrate with two practical examples how these probabilistic propagation algorithms can be applied to high dimensional processes whose conditional independence structure, as well as their underlying distributions, are augmented through the passage of time. © 1999 Elsevier Science B.V. All rights reserved.

Original language	English
Pages (from-to)	99-124
Number of pages	25
Journal	Artificial Intelligence
Volume	107
Issue number	1
Publication status	Published - Jan 1999

Keywords

Dynamic models
Hellingen metric
Influence diagrams
Junction trees
Multivariate state space models
Probabilistic expert systems

Cite this

@article{c80e64ee046545868de6de6cb0a5332e,

title = "Fast Bayes and the dynamic junction forest",

abstract = "It has been shown that junction tree algorithms can provide a quick and efficient method for propagating probabilities in complex multivariate problems when they can be described by a fixed conditional independence structure. In this paper we formalise and illustrate with two practical examples how these probabilistic propagation algorithms can be applied to high dimensional processes whose conditional independence structure, as well as their underlying distributions, are augmented through the passage of time. {\textcopyright} 1999 Elsevier Science B.V. All rights reserved.",

keywords = "Dynamic models, Hellingen metric, Influence diagrams, Junction trees, Multivariate state space models, Probabilistic expert systems",

author = "Smith, {J. Q.} and Papamichail, {K. N.}",

year = "1999",

month = jan,

language = "English",

volume = "107",

pages = "99--124",

journal = "Artificial Intelligence",

issn = "0004-3702",

publisher = "Elsevier BV",

number = "1",

}

TY - JOUR

T1 - Fast Bayes and the dynamic junction forest

AU - Smith, J. Q.

AU - Papamichail, K. N.

PY - 1999/1

Y1 - 1999/1

N2 - It has been shown that junction tree algorithms can provide a quick and efficient method for propagating probabilities in complex multivariate problems when they can be described by a fixed conditional independence structure. In this paper we formalise and illustrate with two practical examples how these probabilistic propagation algorithms can be applied to high dimensional processes whose conditional independence structure, as well as their underlying distributions, are augmented through the passage of time. © 1999 Elsevier Science B.V. All rights reserved.

AB - It has been shown that junction tree algorithms can provide a quick and efficient method for propagating probabilities in complex multivariate problems when they can be described by a fixed conditional independence structure. In this paper we formalise and illustrate with two practical examples how these probabilistic propagation algorithms can be applied to high dimensional processes whose conditional independence structure, as well as their underlying distributions, are augmented through the passage of time. © 1999 Elsevier Science B.V. All rights reserved.

KW - Dynamic models

KW - Hellingen metric

KW - Influence diagrams

KW - Junction trees

KW - Multivariate state space models

KW - Probabilistic expert systems

M3 - Article

SN - 0004-3702

VL - 107

SP - 99

EP - 124

JO - Artificial Intelligence

JF - Artificial Intelligence

IS - 1

ER -

Fast Bayes and the dynamic junction forest

Abstract

Keywords

Fingerprint

Cite this