Thank you for your help. I solved the problem by scrubbing the
graph first with
simplify(as.undirected(graph))
27.10.18 11:39 Szabolcs Horvát
kirjutas:
Actually, I believe that this is a bug. I
reported it a while ago (but I could not recall this
immediately).
I do not consider it a bug in the C interface function,
as when programming in C, it is typically the user who is
expected to supply valid input (and it is not the function
that is expected to check validity).
But in R it is definitely a bug. A high-level interface
should just not accept invalid input.
This happens because this is not a simple
graph. It has multi-edges. The "average" computation
ignores these, the "local" computation does not. I did not
verify if the "local" computation gives reasonable results
for multigraphs, though.
I added
to the attachment the graph that gives me these results.
It is
saved in edgelist format.
27.10.18 10:42 Gábor Csárdi kirjutas:
> Please provide a reproducible example.
>
> Gabor
> On Sat, Oct 27, 2018 at 7:13 AM Raigo Aljand <address@hidden> wrote:
>> Hello,
>>
>> I have a simple question. I looked at the
archives, but it doesn't seem
>> anyone has asked this nor does it seem to be
written in the documentation.
>>
>> Why do these two give me different results?
>>
>> transitivity(graph, type = "average")
>> 0.3772343
>> mean(transitivity(graph, type = "local"), na.rm =
TRUE))
>> 0.2014909
>>
>> I understand the difference between global and
local transitivity and
>> what I read from the internet type="average"
should give me the average
>> local transitivity. Then why does the igraph
average give me a different
>> result from when I calculate the average with R.
What is also strange is
>> that type="average" is either no longer or has
never been documented in
>> the official documentation.
>>
>> If it helps, my graph is directed, but as I
understand it, igraph
>> ignores the direction of the edges.
>>
>> My version of igraph is Ubuntu 18.04 package
r-cran-igraph with version
>> 1.1.2 and R is from the package r-base with
version 3.4.4
>>
>> Thank you for your help.
>>
>>
>>
>> _______________________________________________
>> igraph-help mailing list
>> address@hidden
>> https://lists.nongnu.org/mailman/listinfo/igraph-help
> _______________________________________________
> igraph-help mailing list
> address@hidden
> https://lists.nongnu.org/mailman/listinfo/igraph-help
_______________________________________________
igraph-help mailing list
address@hidden
https://lists.nongnu.org/mailman/listinfo/igraph-help
_______________________________________________
igraph-help mailing list
address@hidden
https://lists.nongnu.org/mailman/listinfo/igraph-help
|