Hmmm, it seems that there is a bug in alpha.centrality, if sparse=TRUE and the graph is undirected.
You need to use sparse=FALSE, or, if your graph is big, then convert it to directed: library(igraph) g <- graph.ring(10) alpha.centrality(g, sparse=FALSE) # [1] -1 -1 -1 -1 -1 -1 -1 -1 -1 -1 alpha.centrality(g, sparse=TRUE) # [1] 1 2 3 4 5 6 7 8 9 11 g2 <- as.directed(g, mode="mutual") alpha.centrality(g2, sparse=TRUE) # [1] -1 -1 -1 -1 -1 -1 -1 -1 -1 -1 On Mon, Jul 21, 2014 at 6:20 AM, Mathieu Ferry <[email protected]> wrote: > Dear all, > > I am trying to apply some centrality measures to a network. > > I find that one node has a degree of 11, which is relatively high, > > but when I compute alpha centrality (alpha = 0.5 , exo= 0.5), I find a > measure that is lower than this (1.875). > > I don't understand how this can be possible, whether it is theoretically > possible actually... > > I mean, how can my alpha centrality measure be lower than 11*0.5 = 5.5? > Shouldn't it be 11*0.5 + ... all other paths? > > The network is undirected, g is an adjacency matrix, with 250 nodes. > > This is the command I used: > katz0.5 <- alpha.centrality(g, nodes=V(g), alpha=0.5, loops=FALSE, > exo=0.5, weights=NA,tol=1e-7, sparse=TRUE) > > Thanks for your help ! > > _______________________________________________ > igraph-help mailing list > [email protected] > https://lists.nongnu.org/mailman/listinfo/igraph-help > _______________________________________________ igraph-help mailing list [email protected] https://lists.nongnu.org/mailman/listinfo/igraph-help
