[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [igraph] dendogram visualization

From: Tamas Nepusz
Subject: Re: [igraph] dendogram visualization
Date: Wed, 5 Nov 2008 15:23:17 +0000

Hi Kurt,

Hmmmm... actually I never tried to draw such a large dendrogram and I'm not really sure if it's worth it or not ;) Your error message is most probably unrelated to ARPACK at all, it might happen that the Python interpreter simply ran out of memory. (Hmmm... any chance of serialising that dendrogram and sending it to me to check?)

The plot() function also supports VertexDendrogram objects, so you can try to plot it to a PNG file, but as said above, I never tried it for such large dendrograms and it might not work at all or it might look ugly.

1 - are there any resources or examples for creating a nice dendrogram viz from an igraph VertexClustering object?
Another program that is capable of drawing dendrograms is DRAWGRAM in Phylip (http://evolution.genetics.washington.edu/phylip/progs.algs.tree.html ), but you have to convert igraph's dendrogram representation to Phylip's format first. There might be other resources out there on the Internet.

2 - does it really make sense to draw a dendrogram from a fastgreedy community structure? unless i misunderstand, the process is agglomerative and terminates before finding one large community.
igraph continues all the way to the top until a single large community remains, then selects the cut level in the dendrogram in a way that maximises modularity; however, the whole dendrogram is computed, so it shouldn't be a problem to plot the whole dendrogram.


On 2008.11.05., at 15:05, Kurt J wrote:

Hi iGraphers,

I'm revisiting some work I've done a while ago and I want to create some visualizations of a community structure dendogram. Specifically, i'm using python and applied community_fastgreedy and community_walktrap to a network w/ just under 20k nodes.

running in the interpreter:

>>> GM.vcG.summary.__doc__
'Draws the dendrogram of the hierarchical clustering in a string'

>>> GM.vcG.summary()
Python(7152,0xa000d000) malloc: *** vm_allocate(size=417792) failed (error code=3)
Python(7152,0xa000d000) malloc: *** error: can't allocate region
Python(7152,0xa000d000) malloc: *** set a breakpoint in szone_error to debug
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
File "build/bdist.macosx-10.3-fat/egg/igraph/clustering.py", line 460, in summary

Ahh malloc my old friend. I know there is the option to change ARPACK options and this might help. But I was keen to get something a bit sexier than an ascii based dendrogram drawing anyway ;)

So let me distill this meandering post:

Kurt J
igraph-help mailing list

reply via email to

[Prev in Thread] Current Thread [Next in Thread]