Bayesian Networks in the Absence of Temporal Information - Bayesian Networks in R: With Applications in Systems Biology - page 28

Original Graphical Model (UG) Bayesian Network (DAG) Equivalence Class (CPDAG)

S TAT S TAT STAT

ANL ANL ANL

ALG ALG ALG

MECH MECH MECH

VECT VECT VECT

Fig. 2.2 The undirected graphical model for the marks data set from Edwards ( 2000 )and Whit-
taker ( 1990 )( left ), the Bayesian network learned from the same data ( center ), and the CPDAG of
its equivalence class ( right )

average branching factor: 1.20

generation algorithm: Empty

Rather than typing the arcs of the network in a two-column format, we can also
create ug starting from an adjacency matrix as follows:

> mat = matrix(c(0, 1, 1, 0, 0, 0, 0, 1, 0, 0, 0, 0,
+ 0,1,1,0,0,0,0,1,0,0,0,0,0),

+ nrow = 5,

+ dimnames = list(nodes(dag), nodes(dag)))

> mat

MECH VECT ALG ANL STAT
MECH 0 0 0 0 0
VECT 1 0 0 0 0
ALG 1 1 0 0 0
ANL 0 0 1 0 0
STAT 0 0 1 1 0
> dag2 = empty.graph(nodes(dag))
> amat(dag2) = mat
> all.equal(dag, dag2)
[1] TRUE

On the other hand, sometimes we may just want to create a new bn object by modi-
fying an existing one. The most straightforward way to accomplish this is by adding
( set.arc ), dropping ( drop.arc ), or reversing ( rev.arc ) arcs in the original
network.