Re: Betweenness Centrality on Directed Graphs

[Suhas Bhairav <suhasbhairav AT gmail.com>]

Hi,

Going by the definition of Betweenness Centrality for directed graphs, the betweenness centrality value obtained for each node from the current implementation of graphstream should be multiplied by 2 in case of a directed graph to fetch the correct betweenness of the respective node. This will give us the appropriate betweenness centrality value for each node in a directed graph. But for an undirected graph, the existing implementation in graphstream is correct. Have I understood it correctly?

Wikipedia excerpt: The division is done by  for directed graphs and  for undirected graphs, where  is the number of nodes in the giant component.

Source: https://en.wikipedia.org/wiki/Betweenness_centrality

Regards

Suhas Bhairav

On 16 June 2016 at 11:04, Suhas Bhairav <suhasbhairav AT gmail.com> wrote:

Hi,

In the documentation related to Betweenness Centrality computation it has been mentioned "This algorithm does not take into account edge direction yet" under the "Usage" section . So does that mean a directed graph is treated as an undirected graph during the  calculation of betweenness centrality value of each node?

If this is the case, we would be probably have to implement our own version of Betweenness Centrality calculation on a directed graph while calculating centrality of each node. Is my understanding correct?

Link : http://graphstream-project.org/doc/Algorithms/Betweenness-Centrality/

Regards

Suhas Bhairav