igraph-help
[Top][All Lists]

## Re: [igraph] Question about degree.distribution

 From: Tamás Nepusz Subject: Re: [igraph] Question about degree.distribution Date: Wed, 8 Jun 2011 20:23:31 +0200

Hi Narcis,

What I don't understand is, why the total degree distribution is not changing (lets say for degrees around 100)
Most of your vertices except one have total degrees larger than 121, so igraph is correct; the cumulative total degree distribution should be almost 1 up to x=122.

Shouldn't the "in" and "out"degree  distribution sum up to the total degree distribution?
No, in general not. The cumulative total degree distribution gives you the number of nodes that have a total degree >= x for any given x. Since there are many ways for a node to have a total degree >= x (e.g., in-degree >= 1 and out-degree >= x-1, or in-degree >= 2 and out-degree >= x-2  and so on), the total cumulative degree distribution is not simply the pointwise sum of the in- and out-degree distributions; I guess it's more like the convolution of the two, but I'm not sure.

--
T.