gRain (version 1.3-2)

grain-main: Graphical Independence Network

Description

Creating grain objects (graphical independence network).

Usage

grain(x, control = list(), smooth = 0, details = 0, data = NULL, ...)
# S3 method for cpt_spec
grain(x, control = list(), smooth = 0, details = 0, ...)
# S3 method for pot_spec
grain(x, control = list(), smooth = 0, details = 0, ...)
# S3 method for pot_rep
grain(x, ...)
# S3 method for cpt_rep
grain(x, ...)
# S3 method for graphNEL
grain(x, control = list(), smooth = 0, details = 0, data = NULL, ...)
# S3 method for dModel
grain(x, control = list(), smooth = 0, details = 0, data = NULL, ...)

Arguments

x

An argument to build an independence network from. Typically a list of conditional probability tables, a DAG or an undirected graph. In the two latter cases, data must also be provided.

control

A list defining controls, see 'details' below.

smooth

A (usually small) number to add to the counts of a table if the grain is built from a graph plus a dataset.

details

Debugging information.

data

An optional data set (currently must be an array/table)

...

Additional arguments, currently not used.

Value

An object of class "grain"

Details

If 'smooth' is non-zero then entries of 'values' which a zero are replaced by the value of 'smooth' - BEFORE any normalization takes place.

References

S<U+00F8>ren H<U+00F8>jsgaard (2012). Graphical Independence Networks with the gRain Package for R. Journal of Statistical Software, 46(10), 1-26. http://www.jstatsoft.org/v46/i10/.

See Also

cptable, compile.grain, propagate.grain, setFinding, setEvidence, getFinding, pFinding, retractFinding

Examples

# NOT RUN {
## Asia (chest clinic) example:
yn   <- c("yes","no")
a    <- cptable(~asia,              values=c(1,99), levels=yn)
t.a  <- cptable(~tub+asia,          values=c(5,95,1,99), levels=yn)
s    <- cptable(~smoke,             values=c(5,5), levels=yn)
l.s  <- cptable(~lung+smoke,        values=c(1,9,1,99), levels=yn)
b.s  <- cptable(~bronc+smoke,       values=c(6,4,3,7), levels=yn)
e.lt <- cptable(~either+lung+tub,   values=c(1,0,1,0,1,0,0,1), levels=yn)
x.e  <- cptable(~xray+either,       values=c(98,2,5,95), levels=yn)
d.be <- cptable(~dysp+bronc+either, values=c(9,1,7,3,8,2,1,9), levels=yn)
plist <- compileCPT(list(a, t.a, s, l.s, b.s, e.lt, x.e, d.be))
bn    <- grain(plist)
bn
summary(bn)
plot(bn)
bnc <- compile(bn, propagate=TRUE)

## If we want to query the joint distribution of the disease nodes,
## computations can be speeded up by forcing these nodes to be in
## the same clique of the junction tree:

bnc2 <- compile(bn, root=c("lung", "bronc", "tub"), propagate=TRUE)

# }
# NOT RUN {
if (require(microbenchmark)){
microbenchmark(
  querygrain(bnc, nodes=c("lung","bronc", "tub"), type="joint"),
  querygrain(bnc2, nodes=c("lung","bronc", "tub"), type="joint")
)}
# }
# NOT RUN {

## Simple example - one clique only in triangulated graph:
plist.s <- compileCPT(list(a, t.a))
bn.s <- grain(plist.s)
querygrain(bn.s)

## Simple example - disconnected network:
plist.d <- compileCPT(list(a, t.a, s))
bn.d <- grain(plist.d)
querygrain(bn.d)

## Create network from data and graph specification.
## There are different ways:

data(HairEyeColor)
hec <- HairEyeColor
daG <- dag(~Hair + Eye:Hair + Sex:Hair)
class(daG)
uG <- ug( ~Eye:Hair + Sex:Hair)
class(uG)

## Create directly from dag:
bn.dag  <- grain(daG, data=hec)
class(bn.dag)
compile(bn.dag)

## Build model from undirected (decomposable) graph
bn.ug  <- grain(uG, data=hec)
class(bn.ug)
compile(bn.ug)

# }