Re: [igraph] Differences between UCINET and igraph for Freeman's degree centralization

[Gábor Csárdi]

On Wed, Mar 27, 2013 at 12:00 AM, Gábor Csárdi <address@hidden> wrote:

On Tue, Mar 26, 2013 at 11:49 PM, Jeremy B. <address@hidden> wrote:

[...]

As a side note, I created fictitious star network data and got 100% in UCINET but 75% in igraph.

An in-star is 100% centralized:

library(igraph)

g <- graph.star(10, mode="in")

centralization.degree(g, mode="in", loops=FALSE)

For the record, the problem was the omission of the 'loops=FALSE' argument, which should be default, I think, as most graphs don't allow loops.

Gabor

[...]