Decomposing Bayesian networks: Triangulation of the moral graph with genetic algorithms

P. Larrañaga; Cindy Kuijpers; M. Poza; R.H. Murga

doi:10.1023/A:1018553211613

Decomposing Bayesian networks: Triangulation of the moral graph with genetic algorithms

P. Larrañaga, Cindy Kuijpers, M. Poza, R.H. Murga

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

61 Citations (Scopus)

Abstract

In this paper we consider the optimal decomposition of Bayesian networks. More concretely, we examine empirically the applicability of genetic algorithms to the problem of the triangulation of moral graphs. This problem constitutes the only difficult step in the evidence propagation algorithm of Lauritzen and Spiegelhalter (1988) and is known to be NP-hard (Wen, 1991). We carry out experiments with distinct crossover and mutation operators and with different population sizes, mutation rates and selection biasses. The results are analysed statistically. They turn out to improve the results obtained with most other known triangulation methods (Kjærulff, 1990) and are comparable to results obtained with simulated annealing (Kjærulff, 1990; Kjærulff, 1992).

Original language	English
Pages (from-to)	19-34
Journal	Statistics and Computing
Volume	7
Issue number	1
DOIs	https://doi.org/10.1023/A:1018553211613
Publication status	Published - Mar 1997
Externally published	Yes

Access to Document

10.1023/A:1018553211613

Fingerprint Dive into the research topics of 'Decomposing Bayesian networks: Triangulation of the moral graph with genetic algorithms'. Together they form a unique fingerprint.

Cite this

@article{24566bd6ebd8439dacad4ab8174b5ecf,

title = "Decomposing Bayesian networks: Triangulation of the moral graph with genetic algorithms",

abstract = "In this paper we consider the optimal decomposition of Bayesian networks. More concretely, we examine empirically the applicability of genetic algorithms to the problem of the triangulation of moral graphs. This problem constitutes the only difficult step in the evidence propagation algorithm of Lauritzen and Spiegelhalter (1988) and is known to be NP-hard (Wen, 1991). We carry out experiments with distinct crossover and mutation operators and with different population sizes, mutation rates and selection biasses. The results are analysed statistically. They turn out to improve the results obtained with most other known triangulation methods (Kj{\ae}rulff, 1990) and are comparable to results obtained with simulated annealing (Kj{\ae}rulff, 1990; Kj{\ae}rulff, 1992).",

author = "P. Larra{\~n}aga and Cindy Kuijpers and M. Poza and R.H. Murga",

year = "1997",

month = mar,

doi = "10.1023/A:1018553211613",

language = "English",

volume = "7",

pages = "19--34",

journal = "Statistics and Computing",

issn = "0960-3174",

publisher = "Springer Netherlands",

number = "1",

}

TY - JOUR

T1 - Decomposing Bayesian networks

T2 - Triangulation of the moral graph with genetic algorithms

AU - Larrañaga, P.

AU - Kuijpers, Cindy

AU - Poza, M.

AU - Murga, R.H.

PY - 1997/3

Y1 - 1997/3

N2 - In this paper we consider the optimal decomposition of Bayesian networks. More concretely, we examine empirically the applicability of genetic algorithms to the problem of the triangulation of moral graphs. This problem constitutes the only difficult step in the evidence propagation algorithm of Lauritzen and Spiegelhalter (1988) and is known to be NP-hard (Wen, 1991). We carry out experiments with distinct crossover and mutation operators and with different population sizes, mutation rates and selection biasses. The results are analysed statistically. They turn out to improve the results obtained with most other known triangulation methods (Kjærulff, 1990) and are comparable to results obtained with simulated annealing (Kjærulff, 1990; Kjærulff, 1992).

AB - In this paper we consider the optimal decomposition of Bayesian networks. More concretely, we examine empirically the applicability of genetic algorithms to the problem of the triangulation of moral graphs. This problem constitutes the only difficult step in the evidence propagation algorithm of Lauritzen and Spiegelhalter (1988) and is known to be NP-hard (Wen, 1991). We carry out experiments with distinct crossover and mutation operators and with different population sizes, mutation rates and selection biasses. The results are analysed statistically. They turn out to improve the results obtained with most other known triangulation methods (Kjærulff, 1990) and are comparable to results obtained with simulated annealing (Kjærulff, 1990; Kjærulff, 1992).

U2 - 10.1023/A:1018553211613

DO - 10.1023/A:1018553211613

M3 - Article

VL - 7

SP - 19

EP - 34

JO - Statistics and Computing

JF - Statistics and Computing

SN - 0960-3174

IS - 1

ER -