[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [igraph] correlation between degree of a node and average degree of
From: |
Simone Gabbriellini |
Subject: |
Re: [igraph] correlation between degree of a node and average degree of its neighbors |
Date: |
Thu, 25 Nov 2010 14:40:31 +0100 |
Dear Jordi,
thanks for sharing your thoughts, but according to Latapy et al., Social
Networks, 30 (2008) 31-48, the correlation has a pretty interesting
interpretation, which does not rely on normalization. The actors-movies network
data they show is an interesting example.
thanks,
Simone
Il giorno 25/nov/2010, alle ore 12.42, jordi torrents ha scritto:
> Hello,
>
> 2010/11/24 Simone Gabbriellini <address@hidden>:
>> thanks,
>>
>> that function is really it... but I am wondering if it works in a bipartite
>> case... it should, right?
>
> You have to be careful in order to make sense of the correlation
> between node degree and average degree of its neighbors in a bipartite
> network. Notice that since there are two sets of nodes and edges only
> can link nodes of different sets, if there is a big difference between
> the number of nodes of the two sets, the correlation will be biased
> because the set that contains less nodes will have higher degrees on
> average.
>
> One possible approach would be to normalize the degree before running
> the correlation. In order to do that in a bipartite network, you will
> have to apply two separate normalizations for each node set:
>
> nd_i = d_i / n_2 for i in V_1
>
> nd_j = d_j / n_1 for j in V_2
>
> where nd_i is the normalized degree for node i (which belongs to node
> set V_1), d_i is the raw degree of that node and n_2 is the number of
> nodes of node set V_2 (which is the maximum degree that a node of set
> V_1 could have).
>
> Hope that helps.
> Salut!
>
> _______________________________________________
> igraph-help mailing list
> address@hidden
> http://lists.nongnu.org/mailman/listinfo/igraph-help