0

I have a very large network with 50,000 nodes. It's sparse. I want to find communities in it, using R. How do I do so? Thank you!

(I tried using igraph, but it won't work because the adjacency matrix is too large.)

The dataset looks like this:

1 0 0 1 0 0 0 1
0 1 0 1 0 1 1 0

...

It's 50,000 x 80.

I found the correlation between every row, creating a correlation matrix that looks like this:

0.14 0.26 0.36
0.24 0.79 0.36

...

It's 50,000 x 50,000.

Then I put it into igraph:

output2<-matrix(ifelse(runif(50000*80)<0.2,1,0),50000,80) # random binary sparse matrix
x2 <- graph.adjacency(cor(t(output2)), weighted=TRUE, diag=FALSE)
x2 = delete.vertices(x2,which(degree(x)<1))
x2 = as.undirected(x2)
b2 <- walktrap.community(x2)
k3<-groups(b2)

However, igraph says it can't create the adjacency graph, as it's too large.

Hallo
  • 142
  • 10
  • 1
    To whoever downvoted my comment, please let me know how I can improve my question! – Hallo Jun 15 '17 at 13:23
  • 1
    I didn't downvote, but please read the info about [how to ask a good question](http://stackoverflow.com/help/how-to-ask) and how to give a [reproducible example](http://stackoverflow.com/questions/5963269) . This will make it much easier for others to help you. – Jaap Jun 15 '17 at 13:25
  • 1
    That would be me. The main reason I downvoted is because while you have the "informed" badge, meaning that you read the rules of this site, you still posted a question that is outside the scope of stack overflow – Sotos Jun 15 '17 at 13:25
  • Oh. Can this be migrated? – Hallo Jun 15 '17 at 13:26
  • Migrate where? I suggest you do some research, come up with a code and IF you have a specific question about your code, then ask a question. – Sotos Jun 15 '17 at 13:27
  • I just put in the code – Hallo Jun 15 '17 at 13:29
  • Much better. I will retract my votes but still I urge you to look at the links @Jaap posted and provide reproducible examples. – Sotos Jun 15 '17 at 13:31
  • Will do. Thank you! – Hallo Jun 15 '17 at 13:36

0 Answers0