qpRndGraph {qpgraph}R Documentation

Random undirected graphs with maximum connectivity degree

Description

Builds a random undirected graph with a bounded maximum connectivity degree (boundary) on every vertex.

Usage

qpRndGraph(n.vtx, n.bd)

Arguments

n.vtx number of vertices.
n.bd maximum boundary for every vertex.

Details

This is a very simple function to generate random undirected graphs where we impose a maximum order of correlation between disconnected vertices when using it to sample multivariate normal data reflecting the conditional independencies encoded in this graph. Note that the maximum order of correlation between two disconnected vertices is bounded by the minimum degree of connectivity of the two vertices.

The algorithm employed is not designed to enforce a uniform probability distribution on the set of graphs with the given maximum boundary that may be sampled with positive probability.

Value

The incidence matrix of the resulting graph.

Author(s)

R. Castelo and A. Roverato

References

Castelo, R. and Roverato, A. A robust procedure for Gaussian graphical model search from microarray data with p larger than n, J. Mach. Learn. Res., 7:2621-2650, 2006.

See Also

qpSampleMvnorm

Examples

nVertices <- 50 # number of vertices
maxCon <- 5  # maximum connectivity per vertex

I <- qpRndGraph(n.vtx=nVertices, n.bd=maxCon)

summary(rowSums(I))

[Package qpgraph version 1.0.0 Index]