gmodel.P function

Generate graphs given a probability matrix

Given an $(n\times n)$ probability matrix $P$, gmodel.P generates binary observation graphs corresponding to Bernoulli distribution whose parameter matches to the element of $P$.

gmodel.P(P, rep = 1, noloop = TRUE, symmetric.out = FALSE)

Arguments

• P: an $(n\times n)$ probability matrix.
• rep: the number of observations to be generated.
• noloop: a logical value; TRUE for graphs without self-loops, FALSE otherwise.
• symmetric.out: a logical value; FALSE for generated graphs to be nonsymmetric, TRUE otherwise. Note that TRUE is supported only if the input matrix P is symmetric.

Returns

depending on rep value, either

• (rep=1): an (n-by-n) observation matrix, or
• (rep>1): a length-rep list where each element is an observation is an (n-by-n) realization from the model.

Examples

## set inputs
modelP <- matrix(runif(16),nrow=4)

## generate 3 observations without self-loops.
out <- gmodel.P(modelP,rep=3,noloop=TRUE)

## visualize generated graphs
opar = par(no.readonly=TRUE)
par(mfrow=c(1,3), pty="s")
image(out[[1]], main="1st sample")
image(out[[2]], main="2nd sample")
image(out[[3]], main="3rd sample")
par(opar)