Re: [igraph] Questions about decompose.graph?

[Tamas Nepusz]

> When I call decompose.graph, which can seperate a graph into components, I
> have two questions:
> 
> 1. What's the seperation rules? Or what measure is used in a graph
> decomposition?
decompose.graph simply returns the weakly or strongly connected components
of a graph; i.e., if vertex A is reachable from vertex B and vice versa,
then A and B will be in the same component.

> 2. Please see the following graph drawing, each component is drawn in a
> color as the backgrounds. Who can tell me how can do this?
igraph includes several non-overlapping community detection functions that
can decompose graphs into the modules you show on the picture. Take a look
at these help pages in R:

?edge.betweenness.community
?fastgreedy.community
?leading.eigenvector.community
?spinglass.community
?walktrap.community

However, the figure you showed contains *overlapping* communities, and none
of the above functions actually gives you overlapping ones. If you need an
overlapping community detection method, it only takes a few lines of code to
implement it:

http://igraph.wikidot.com/community-detection-in-r#toc0

-- 
T.