-1

I have a network with 68,254 clusters in it. I know that the largest cluster has 900k vertices and ~1M edges. And simply by manually looking at other clusters, they seem to have 2 or 3 connected vertices.

Is there a way in R to actually tell you the descriptives of these clusters (preferably excluding the largest one), like average number of vertices, edges, in-, out-degree distribution, density, etc.?

It's a directed graph. I'm using the igraph module for the analysis.

Also, is there a way to sort, say, first 100 clusters decreasing by their vertex count? I'm pretty much a novice with R. Thanks.

Zlo
  • 1,150
  • 2
  • 18
  • 38
  • It would be easier to help you if you included a specific [reproducible example](http://stackoverflow.com/questions/5963269/how-to-make-a-great-r-reproducible-example). Show the code you've tried so far. It's also best if you stick to one question at a time since only one accepted answer is allowed. – MrFlick Nov 26 '14 at 19:08

1 Answers1

0

I don't know if there is a special function in R to do this. But it's very simple with other packages. You have the network metrics on each of your nodes, each of your cluster should have an ID. You merge this cluster ID on the nodes, and you compute this average number for each metrics. If you are a novice i R, you can even do that with RcmdR.

Bérengère
  • 21
  • 4