Re: [igraph] issue with evcent

[Gábor Csárdi]

Yeah, but what does that mean? Vertex '1' is the most central, just
because it has a self loop? I think this is not normally the way you
want a centrality measure work.

Although there might be some exceptions of course. (Even if I can't
really think of any.) So, yeah, this can be a bug. I don't see any
trivial solution to it right now. We do nothing special about loop
edges I think, so I would expect it to work well, but apparently it
does not. I submitted a bug report:
https://bugs.launchpad.net/igraph/+bug/360649

Thanks, Best,
Gabor

On Mon, Apr 13, 2009 at 8:54 PM, Matthew Galati <address@hidden> wrote:
> Why can't we have self-loops for eigenvector centrality?
>
> For example, a valid eigenvector/value pair for this example could be:
>  vector = 0.8019,1,0.44504
>  value  = 1.80194
>
>
>> -----Original Message-----
>> From: address@hidden
>> [mailto:address@hidden On
>> Behalf Of Gábor Csárdi
>> Sent: Monday, April 13, 2009 2:46 PM
>> To: Help for igraph users
>> Subject: Re: [igraph] issue with evcent
>>
>> Hi Matt,
>>
>> I am not sure that eigenvector centrality makes sense at all for
>> graphs with self-loops. Maybe you just want to remove the self-loops
>> with 'simplify()'.
>>
>> Best,
>> Gabor
>>
>> On Mon, Apr 13, 2009 at 8:34 PM, Matthew Galati
>> <address@hidden> wrote:
>> > Hi - for this tiny example graph, I am getting an eigenvector/value
>> pair that does not seem valid.
>> >
>> >> g<-read.graph("xxx",format="edgelist")
>> >> gu<-as.undirected(g)
>> >
>> >> print.igraph(gu)
>> > Vertices: 3
>> > Edges: 3
>> > Directed: FALSE
>> > Edges:
>> >
>> > [0] 0 -- 1
>> > [1] 0 -- 2
>> > [2] 1 -- 1
>> >
>> >> evcent(gu)
>> > $vector
>> > [1] 0.4811943 1.0000000 0.1939366
>> >
>> > $value
>> > [1] 2.481194
>> >
>> > The eigen pair is suppose to satisfy A.x=lambda.x
>> >
>> > But, in the second component, we have 1.48 != 2.48
>> >
>> > Note, I checked some other examples where there are no self-loops and
>> they seem ok. So, the issue might be related to self-loops.
>> >
>> > Any ideas?
>> >
>> > Thanks,
>> > Matt
>> >
>> >
>> >
>> >
>> >
>> > _______________________________________________
>> > igraph-help mailing list
>> > address@hidden
>> > http://lists.nongnu.org/mailman/listinfo/igraph-help
>> >
>>
>>
>>
>> --
>> Gabor Csardi <address@hidden>     UNIL DGM
>>
>>
>> _______________________________________________
>> igraph-help mailing list
>> address@hidden
>> http://lists.nongnu.org/mailman/listinfo/igraph-help
>
>
>
> _______________________________________________
> igraph-help mailing list
> address@hidden
> http://lists.nongnu.org/mailman/listinfo/igraph-help
>



-- 
Gabor Csardi <address@hidden>     UNIL DGM