Number of the records: 1

Triangulation Heuristics for BN2O Networks

SYSNO ASEP	0327312
Document Type	C - Proceedings Paper (int. conf.)
R&D Document Type	Conference Paper
Title	Triangulation Heuristics for BN2O Networks
Title	Heuristiky pro triangulaci sítí typu BN2O
Author(s)	Savický, Petr (UIVT-O)_{SAI, RID, ORCID} Vomlel, Jiří (UTIA-B)_{RID, ORCID}
Source Title	Symbolic and Quantitative Approaches to Reasoning with Uncertainty. - Berlin : Springer, 2009 / Sossai C. ; Chemello G. - ISSN 0302-9743 - ISBN 978-3-642-02905-9
Pages	s. 566-577
Number of pages	12 s.
Action	ECSQARU 2009. European Conference /10./
Event date	01.07.2009-03.07. 2009
VEvent location	Verona
Country	IT - Italy
Event type	WRD
Language	eng - English
Country	DE - Germany
Keywords	Bayesian network ; BN2O ; noisy-or ; graphical transformation ; parent divorcing ; tensor rank-one decomposition
Subject RIV	BA - General Mathematics
R&D Projects	1M0545 GA MŠMT - Ministry of Education, Youth and Sports (MEYS)
	1ET100300517 GA AV ČR - Academy of Sciences of the Czech Republic (AV ČR)
	1M0572 GA MŠMT - Ministry of Education, Youth and Sports (MEYS)
	GEICC/08/E010 GA ČR - Czech Science Foundation (CSF)
	GA201/09/1891 GA ČR - Czech Science Foundation (CSF)
CEZ	AV0Z10300504 - UIVT-O (2005-2011)
	AV0Z10750506 - UTIA-B (2005-2011)
UT WOS	000268585700049
EID SCOPUS	69049089935
DOI	10.1007/978-3-642-02906-6_49
Annotation	A BN2O network is a Bayesian network having the structure of a bipartite graph with all edges directed from one part (the top level) toward the other (the bottom level) and where all conditional probability tables are noisy-or gates. In order to perform efficient inference, graphical transformations of these networks are performed. The complexity of inference is proportional to the total table size of tables corresponding to the cliques of the triangulated graph. Therefore in order to get efficient inference it is desirable to have small cliques in the triangulated graph. We analyze existing heuristic triangulation methods applicable to BN2O networks after transformations using parent divorcing and tensor rank-one decomposition and suggest several modifications. Both theoretical and experimental results confirm that tensor rank-one decomposition yields better results than parent divorcing in randomly generated BN2O networks that we tested.
Workplace	Institute of Computer Science
Contact	Tereza Šírová, sirova@cs.cas.cz, Tel.: 266 053 800
Year of Publishing	2010

Number of the records: 1