read.graph: Read and write a graph-object

In INBO-BMK/INLA: Full Bayesian Analysis of Latent Gaussian Models using Integrated Nested Laplace Approximations

Description

Construct a graph-object from a file or a matrix; write graph-object to file

Usage

inla.read.graph(..., size.only = FALSE)
inla.write.graph(graph, filename = "graph.dat", mode = c("binary", "ascii"), ...)

## S3 method for class 'inla.graph'
summary(object, ...)
## S3 method for class 'inla.graph'
plot(x, y, ...)
## S3 method for class 'inla.graph.summary'
print(x, ...)

Arguments

filename	The filename of the graph.
graph	An inla.graph-object, a (sparse) symmetric matrix, a filename containing the graph, a list or collection of characters and/or numbers defining the graph, or a neighbours list with class nb (see spdep::card and spdep::poly2nb for for details of nb and an example a function returning an nb object
mode	The mode of the file; ascii-file or a (gzip-compressed) binary. Default value depends on the inla.option internal.binary.mode which is default TRUE; see inla.setOption.
object	An inla.graph -object
x	An inla.graph -object
y	Not used
size.only	Only read the size of the graph
...	Additional arguments. In inla.read.graph, then it is the graph definition (object, matrix, character, filename), plus extra arguments. In inla.write.graph it is extra arguments to inla.read.graph.

Value

The output of inla.read.graph, is an inla.graph object, with elements

n	is the size of the graph
nnbs	is a vector with the number of neigbours
nbs	is a list-list with the neigbours
cc	list with connected component information idis a vector with the connected component id for each node (starting from 1) nis the number of connected components nodesis a list-list of nodes belonging to each connected component

Methods implemented for inla.graph are summary and plot. The method plot require the libraries Rgraphviz and graph from the Bioconductor-project, see https://www.bioconductor.org.

Author(s)

Havard Rue [e-mail]

See Also

inla.spy

Examples

## a graph from a file
cat("3 1 1 2 2 1 1 3 0\n", file="g.dat")
g = inla.read.graph("g.dat")
## writing an inla.graph-object to file
g.file = inla.write.graph(g, mode="binary")
## re-reading it from that file
gg = inla.read.graph(g.file)
summary(g)
##
Not run:
plot(g)
inla.spy(g)
## when defining the graph directly in the call, 
## we can use a mix of character and numbers
g = inla.read.graph(c(3, 1, "1 2 2 1 1 3", 0))
inla.spy(c(3, 1, "1 2 2 1 1 3 0"))
inla.spy(c(3, 1, "1 2 2 1 1 3 0"),  reordering=3:1)
inla.write.graph(c(3, 1, "1 2 2 1 1 3 0"))

## building a graph from adjacency matrix
adjacent = matrix(0, nrow = 4, ncol = 4)
adjacent[1,4] = adjacent[4,1] = 1
adjacent[2,4] = adjacent[4,2] = 1
adjacent[2,3] = adjacent[3,2] = 1
adjacent[3,4] = adjacent[4,3] = 1
g = inla.read.graph(adjacent)
plot(g)
summary(g)
End(Not run)

INBO-BMK/INLA documentation built on Dec. 4, 2019, 11:43 p.m.