Re: [igraph] Question about "similarity.dice()"

[Tamás Nepusz]

Hi,

> I want to get the number of shared neighbors in a node list.
Try cocitation() or bibcoupling() -- they should be the same for an undirected 
graph.

> I think if I multiply "the dice similarity value of two nodes" by "the 
> average degree of the two nodes", I will get  the number of the shared 
> neighbors of the two nodes. Right?
Yes, it seems so.

> BTW, is there anything special for two nodes that connected directly, when we 
> calculate the shared neighbors(i.e., two nodes share some neighbors, and at 
> the same time they connected to each other directly).
No special treatment there; if nodes A and B are connected, then B is included 
in the set of neighbors for A and A is included in the set of neighbors for B. 
(Otherwise you would get zero similarity for two vertices that are connected to 
each other but no one else, which does not make much sense IMHO).

-- 
T.