Maximum Likelihood Learning of Conditional MTE Distributions

Helge Langseth, Thomas Dyhre Nielsen, Rafael Rumí, Antonio Salmerón

Research output: Contribution to journalConference article in JournalResearchpeer-review

13 Citations (Scopus)
305 Downloads (Pure)

Abstract

We describe a procedure for inducing conditional densities within the mixtures of truncated exponentials (MTE) framework. We analyse possible conditional MTE specifications and propose a model selection scheme, based on the BIC score, for partitioning the domain of the conditioning variables. Finally, experimental results demonstrate the
applicability of the learning procedure as well as the expressive power of the conditional MTE distribution.
Original languageEnglish
Book seriesLecture Notes in Computer Science
Volume5590
Pages (from-to)240-251
ISSN0302-9743
DOIs
Publication statusPublished - 2009
EventThe European Conference on Symbolic and Quantitative Approaches to Reasoning with Uncertainty - Verona, Italy
Duration: 1 Jul 20093 Jul 2009
Conference number: 10

Conference

ConferenceThe European Conference on Symbolic and Quantitative Approaches to Reasoning with Uncertainty
Number10
CountryItaly
CityVerona
Period01/07/200903/07/2009

Fingerprint

Truncated Distributions
Exponential distribution
Maximum likelihood
Maximum Likelihood
Conditional Density
Expressive Power
Model Selection
Conditioning
Partitioning
Experimental Results
Demonstrate
Learning

Cite this

Langseth, Helge ; Nielsen, Thomas Dyhre ; Rumí, Rafael ; Salmerón, Antonio. / Maximum Likelihood Learning of Conditional MTE Distributions. In: Lecture Notes in Computer Science. 2009 ; Vol. 5590. pp. 240-251.
@inproceedings{61172bf04b7c11dead10000ea68e967b,
title = "Maximum Likelihood Learning of Conditional MTE Distributions",
abstract = "We describe a procedure for inducing conditional densities within the mixtures of truncated exponentials (MTE) framework. We analyse possible conditional MTE specifications and propose a model selection scheme, based on the BIC score, for partitioning the domain of the conditioning variables. Finally, experimental results demonstrate the applicability of the learning procedure as well as the expressive power of the conditional MTE distribution.",
author = "Helge Langseth and Nielsen, {Thomas Dyhre} and Rafael Rum{\'i} and Antonio Salmer{\'o}n",
note = "Titel: Proceedings of the Tenth European Conference on Symbolic and Quantitative Approaches to Reasoning with Uncertainty Oversat titel: Oversat undertitel: Forlag: Springer ISBN (Trykt): 978-3-642-02905-9 ISBN (Elektronisk): Publikationsserier: Lecture Notes in Computer Science, Springer Verlag, 0302-9743, 1611-3349, 5590",
year = "2009",
doi = "10.1007/978-3-642-02906-6_22",
language = "English",
volume = "5590",
pages = "240--251",
journal = "Lecture Notes in Computer Science",
issn = "0302-9743",
publisher = "Physica-Verlag",

}

Maximum Likelihood Learning of Conditional MTE Distributions. / Langseth, Helge; Nielsen, Thomas Dyhre; Rumí, Rafael; Salmerón, Antonio.

In: Lecture Notes in Computer Science, Vol. 5590, 2009, p. 240-251.

Research output: Contribution to journalConference article in JournalResearchpeer-review

TY - GEN

T1 - Maximum Likelihood Learning of Conditional MTE Distributions

AU - Langseth, Helge

AU - Nielsen, Thomas Dyhre

AU - Rumí, Rafael

AU - Salmerón, Antonio

N1 - Titel: Proceedings of the Tenth European Conference on Symbolic and Quantitative Approaches to Reasoning with Uncertainty Oversat titel: Oversat undertitel: Forlag: Springer ISBN (Trykt): 978-3-642-02905-9 ISBN (Elektronisk): Publikationsserier: Lecture Notes in Computer Science, Springer Verlag, 0302-9743, 1611-3349, 5590

PY - 2009

Y1 - 2009

N2 - We describe a procedure for inducing conditional densities within the mixtures of truncated exponentials (MTE) framework. We analyse possible conditional MTE specifications and propose a model selection scheme, based on the BIC score, for partitioning the domain of the conditioning variables. Finally, experimental results demonstrate the applicability of the learning procedure as well as the expressive power of the conditional MTE distribution.

AB - We describe a procedure for inducing conditional densities within the mixtures of truncated exponentials (MTE) framework. We analyse possible conditional MTE specifications and propose a model selection scheme, based on the BIC score, for partitioning the domain of the conditioning variables. Finally, experimental results demonstrate the applicability of the learning procedure as well as the expressive power of the conditional MTE distribution.

U2 - 10.1007/978-3-642-02906-6_22

DO - 10.1007/978-3-642-02906-6_22

M3 - Conference article in Journal

VL - 5590

SP - 240

EP - 251

JO - Lecture Notes in Computer Science

JF - Lecture Notes in Computer Science

SN - 0302-9743

ER -