plot.nethetclustering: Plot networks

Description

This function takes the output of screen_cv.glasso or mixglasso and creates a network plot using the network library.

Usage

## S3 method for class 'nethetclustering':
plot(x, node.names = rownames(net.clustering$Mu),
  group.names = sort(unique(net.clustering$comp)), p.corrs.thresh = 0.2,
  print.pdf = FALSE, pdf.filename = "networks", ...)

Arguments

A network clustering object as returned by screen_cv.glasso or mixglasso.

node.names

Names for the nodes in the network. If NULL, names from net.clustering will be used.

group.names

Names for the clusters or groups. If NULL, names from net.clustering will be used (by default these are integets 1:numClusters).

p.corrs.thresh

Threshold applied to the absolute partial correlations. Edges that are below the threshold in all of the groups are not displayed.

print.pdf

If TRUE, save the output as a PDF file.

pdf.filename

If print.pdf is TRUE, specifies the file name of the output PDF file.

...

Further arguments

Value

Returns NULL and prints out the networks (or saves them to pdf if print.pdf is TRUE. The networks are displayed as a series of nComps+1 plots, where in the first plot edge widths are shown according to the maximum partial correlation of the edge over all groups. The following plots show the edges for each group. Positive partial correlation edges are shown in black, negative ones in blue. If an edge is below the threshold on the absolute partial correlation, it is displayed in gray or light blue respectively.