igraph-help
[Top][All Lists]

## Re: [igraph] Getting two different values for "clustering coefficient"

 From: Gábor Csárdi Subject: Re: [igraph] Getting two different values for "clustering coefficient" Date: Thu, 8 Jul 2010 15:11:18 +0200

```Hi,

On Thu, Jul 8, 2010 at 3:00 PM, shahab mokarizadeh
> Hi,
>
> I am sure that I am missing something, but when I measure the alpha
> coeffiecency for "out degree" distribution of my graph using following
> command
> power.law.fit(outDegree_, xmin=1 )
> I get  :
> Coefficients:
>   alpha
> 1.168368
> and the following warning messages:
> "
> Warning messages:
> 1: In xy.coords(x, y, xlabel, ylabel, log) :
>  1 x value <= 0 omitted from logarithmic plot
> 2: In xy.coords(x, y, xlabel, ylabel, log) :
>  311 y values <= 0 omitted from logarithmic plot
> 3: In xy.coords(x, y, xlabel, ylabel, log) :
>  1 x value <= 0 omitted from logarithmic plot
> 4: In xy.coords(x, y, xlabel, ylabel, log) :
>  311 y values <= 0 omitted from logarithmic plot
> 5: In power.law.fit(deg, xmin = 10) :
>  too small values eliminated from vector
> 6: In power.law.fit(outDegree_, xmin = 1) :
>  too small values eliminated from vector
> "
> 1-What do these warnings try to show? and why does it happen? do they
> have any effect on final computation?

Warnings 1-4 are not for power.law.fit(), but for the plotting you did
before it, possibly from tkigraph().
Warning 5 is for another power.law.fit() you did before, possibly from
tkigraph() as well.
It just says that 'xmin=10' caused the omission of some values from
the computation.
Warning 6 is the same, for your interactive command.

> 2 -but when I invoke  using  Centrality->Fit a Power Law to Degree
> Distribution using tkigraph tool, it shows the following value:
> exponent : 1.63
> standard error: 0,03

xmin=10 is the default for the version in tkigraph(), you used xmin=1.

Best,
Gabor

> Maybe I am mistaking, but I suppose that they both should show a same value?
>
>
> regards,
> /Shahab
>
> _______________________________________________
> igraph-help mailing list
> http://lists.nongnu.org/mailman/listinfo/igraph-help
>

--