[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[igraph] Re: betweenness and degree correlation

From: Tamas Nepusz
Subject: [igraph] Re: betweenness and degree correlation
Date: Tue, 23 Jun 2009 22:22:55 +0100

Hi Yong,

Are you sure you are using the patched igraph version (the one that uses a long double to count the number of shortest paths)? If so, can you check how many bytes does a long double use on your platform? (It's 16 bytes on mine, both on OS X and on openSuSE). You can use the following program:

#include <stdio.h>
int main(int argc, char* argv[]) {
  printf("%d\n", sizeof(long double));
  return 0;


On 2009.06.23., at 18:55, Yong Zou wrote:

Hi Tamas,

Here is my result: 1189955308309381376.000000, vertex 433

I did in the following way (exactly the same way as we discussed last time):
igraph_read_graph_edgelist(&graph, f_edge, 0, IGRAPH_UNDIRECTED);
igraph_betweenness(&graph, &result, igraph_vss_all(), IGRAPH_UNDIRECTED);

What's wrong here?


On Tue, Jun 23, 2009 at 6:03 PM, Tamas Nepusz<address@hidden> wrote:
Hi Yong,

For the betweenness centrality, it seems very easy to have overflow
now for my other network. Please have a look at my attachment (edge
list), which has only 1000 nodes. But I really want to analyze the BC
for this network.
The betweenness calculation seems fine for me for this network; i.e.,
the maximum betweenness is 3641.928 and the minimum is zero. What are
your results?

Another problem is about the degree-degree correlation, namely, I want
to calculate something P(k'|k) which is so called assortativity
coefficient. But I'm sorry I couldn't find it somewhere from igraph.
This is not implemented directly in igraph, but it's fairly simple (at least if you are thinking about the assortativity coefficient as defined
by Newman). This is calculated in Python as follows:


I guess it's not too complicated to translate this back to C.


reply via email to

[Prev in Thread] Current Thread [Next in Thread]