CODEWITHSUNDEEP
rigraphsocial-networkingsnastatnet
Jonathan D.
Jonathan D.

Reputation: 17

In-Degree Bonacich Power Centrality in R?

I am attempting to identify a method to calculate in-degree Bonacich Power Centrality in R. I'm a long-time UCINET user attempting to make the switch. In UCINET, this is done selecting Beta Centrality (Bonacich Power), and selecting "in-centrality" for the direction.

In R, it doesn't seem as though there is a way to calculate this using either sna or igraph packages. Here it is for bonpow in sna:

bonpow(dat, g=1, nodes=NULL, gmode="digraph", diag=FALSE, tmaxdev=FALSE, 
       exponent=1, rescale=FALSE, tol=1e-07)

I do specify digraph, but I am not able to replicate the analysis in R.

Similarly, here it is for power_centrality in igraph:

power_centrality(graph, nodes = V(graph), loops = FALSE,
      exponent = 1, rescale = FALSE, tol = 1e-07, sparse = TRUE)

Here, there does not seem to be a way to specify that it is a directed graph (although you can specify it when defining the network). However, you can estimate it for betweenness centrality.

In neither case do I seem to be able to specify in-degree or out-degree power centrality. Is there something either in these or in a different package that I may be overlooking?

Upvotes: 0

Views: 465

Answers (1)

gvegayon
gvegayon

Reputation: 853

I'm not sure about what do you mean by direction since the original paper, seems to me, does not deal with it. Now, a thing that is usually done with these statistics that are calculated directly from the adjacency matrix is "change the direction" by taking the transpose of the statistic (for example, when computing exposure in the netdiffuseR package we allow the user to compute "incoming" or "outgoing" exposure by just taking the transpose of the adjacency matrix). When you take the transpose, you are essentially flipping the directionality of the ties, i.e. i->j turns to j->i.

If that's what UCINET does (again, not completely sure what it is), then you can get the "incoming"/"outgoing" version by transposing the network. Here is a toy example:

# Loading the sna package (btw: igraph's implementation is a copy of
# sna's). I wrap it around suppressMessages to avoid the verbose
# print that the package has
suppressMessages(library(sna))

# This is random graph I generated with 10 vertices
net <- structure(
  c(0, 0, 0, 1, 0, 1, 1, 0, 1, 0, 1, 0, 1, 1, 0, 0, 1, 
    0, 1, 1, 0, 0, 0, 0, 0, 1, 1, 0, 0, 0, 1, 0, 0, 0, 0, 0, 1, 1, 
    0, 0, 0, 1, 0, 0, 0, 0, 1, 1, 0, 1, 0, 0, 0, 1, 1, 0, 0, 1, 1, 
    0, 1, 0, 0, 0, 0, 1, 0, 1, 1, 1, 0, 1, 1, 1, 0, 0, 0, 0, 1, 1, 
    0, 0, 0, 1, 1, 0, 0, 1, 0, 1, 0, 1, 1, 1, 1, 1, 0, 1, 0, 0),
  .Dim = c(10L, 10L)
  )

# Here is the default
bonpow(net)
#>  [1] -0.8921521 -0.7658658 -0.9165947 -1.4176664 -0.6151369 -0.7862345
#>  [7] -0.9206684 -1.3565601 -1.0347335 -1.0062173

# Here I'm getting the transpose of the adjmat
net2 <- t(net)

# The output is different (as you can see)
bonpow(net2)
#>  [1] -0.8969158 -1.1026305 -0.6336011 -0.7158869 -1.2960022 -0.9545159
#>  [7] -1.1684592 -0.8845729 -1.0368018 -1.1190876

Created on 2019-11-20 by the reprex package (v0.3.0)

Upvotes: 0

Related Questions