Received: 29 September 2016 Revised: 24 January 2017 Accepted: 7 March 2017
DOI: 10.1111/coin.12130
ORIGINAL ARTICLE
On a simple method for testing independencies in
Bayesian networks
Cory J. Butz
1
André E. dos Santos
1
Jhonatan S. Oliveira
1
Christophe Gonzales
2
1
Department of Computer Science,
University of Regina, Regina, Canada
2
LIP6 - dpartement DESIR, Universit
Pierre et Marie Curie, Paris, France
Correspondence
Cory J. Butz, Department of Computer
Science, University of Regina, Regina,
Saskatchewan, Canada.
Email: Butz@cs.uregina.ca
Funding information
NSERC Discovery, Grant/Award
Number: 238880
Abstract
Testing independencies is a fundamental task in reason-
ing with Bayesian networks (BNs). In practice, d-separation
is often used for this task, since it has linear-time com-
plexity. However, many have had difficulties understanding
d-separation in BNs. An equivalent method that is easier
to understand, called m-separation, transforms the problem
from directed separation in BNs into classical separation in
undirected graphs. Two main steps of this transformation are
pruning the BN and adding undirected edges.
In this paper, we propose u-separation as an even simpler
method for testing independencies in a BN. Our approach also
converts the problem into classical separation in an undirected
graph. However, our method is based upon the novel con-
cepts of inaugural variables and rationalization. Thereby, the
primary advantage of u-separation over m-separation is that
m-separation can prune unnecessarily and add superfluous
edges. Our experiment results show that u-separation performs
73% fewer modifications on average than m-separation.
KEYWORDS
Bayesian networks, d-separation, m-separation, testing independencies
1 INTRODUCTION
Pearl
1
states that perhaps the founding of Bayesian networks (BNs)
2-4
made its greatest impact through
the notion of d-separation. d-Separation
5
is a graphical method for deciding which conditional inde-
pendence relations are implied by the directed acyclic graph (DAG) of a BN. To test whether 2 sets
X and Z of variables are conditionally independent given a third set Y of variables, denoted I(X, Y, Z),
d-separation checks whether every path from X to Z is “blocked” by Y in the DAG. d-Separation
Computational Intelligence. 2017;1–13. wileyonlinelibrary.com/journal/coin © 2017 Wiley Periodicals, Inc. 1