1

I have an edgelist of an undirected graph structure.

data <- data.frame(i = c("var1","var3","var5","var7","var7","var7","var11"),
                    j = c("var2","var4","var6","var8","var9","var10", "var1"))
head(data)

     i     j
1 var1  var2
2 var3  var4
3 var5  var6
4 var7  var8
5 var7  var9
6 var7 var10

I would like to retrieve a list of closed circuits from this dataset. Is there a method e.g. in library(igraph) that would allow you to easily do it?

Extra issue that may be solved with the same method: Incidentally I realized that what I've wanted from the data all along is to just cluster elements in an undirected graph that were very strongly connected. Perhaps a closed circuit method would also solve the problem of finding more connected groups in an undirected graph (perhaps over an arbitrary threshold). Though being new to graph theory, this Wikipedia page on Strength of Graph has something along these lines. The idea is to find natural groups within a set of nodes depending on the structure of their interconnections, with the number of groups depending purely on the data and some nodes possibly belonging to purely monadic groups.

Is there some predefined method or sequence of methods I could do to get it? Thanks!

puslet88
  • 1,288
  • 15
  • 25
  • There is a function in igraph to find the girth (the shortest cycle), but it cannot find all cycles currently. – Gabor Csardi Mar 19 '15 at 18:50
  • Thanks! That's good to know. I guess printing out all circles isn't really a thing often used in network analysis. – puslet88 Mar 20 '15 at 06:41
  • yes, I guess usually there is too many of them even in a smallish network. – Gabor Csardi Mar 20 '15 at 13:21
  • Well, it depends on the data. I mean technically there would be just one closed cycle in an undirected graph network. I guess the first question could be rephrased as well to: print out members of each of the graph networks in a set. Finding strongly connected elements does seem like a function of general use though. – puslet88 Mar 20 '15 at 14:49

0 Answers0