On 24.03.2017 11:18, yerali wrote:
Dear Tiago, Thanks for your answer. Sorry for the delay of my reply but I was learning how to add a property from external data.
I just wanted to answer:
And how do you make this assessment?
So, in this picture the separation into communities using partition stability, as an example: <http://main-discussion-list-for-the-graph-tool-project.982480.n3.nabble.com/file/n4027150/dib_ps.jpg> I think colors well represent set of nodes densely connected, which is important at least in my current case of study, where links represent causality.
This is a perfect example of what I am trying to convey to you. What you are seeing are basically three groups of nodes with low degree around nodes with high in-degree, in addition to some smaller groups. Any random graph with the same degree sequence will admit similar partitions, although they carry no meaning from a generative point of view. You can check this by randomizing your graph with random_rewire() and then obtaining the partition again with the same method. You will probably see a similar division. In other words, these are not statistically significant communities; they arise out of random fluctuations. Best, Tiago -- Tiago de Paula Peixoto <tiago@skewed.de>