[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [igraph] Fastgreedy Algorithm for Community Detection in igraph for
|
From: |
Tamas Nepusz |
|
Subject: |
Re: [igraph] Fastgreedy Algorithm for Community Detection in igraph for RStudio - fatal error. |
|
Date: |
Thu, 29 Nov 2012 09:36:54 +0100 |
|
User-agent: |
Mozilla/5.0 (X11; Linux x86_64; rv:17.0) Gecko/17.0 Thunderbird/17.0 |
> May I ask you one last question? You told me, that the Optimal Modularity
> Algorithm works only for networks with 30-40 nodes. So it's not usable for
> my pretty large network. But do you know how long it would take to compute
> the community structure in this network with 307 nodes and 21'406 edges on a
> server with 4 cores, each 3 GHz, and 8 gb RAM?
Probably a week -- but this is just a wild guess. There is no "rule" to
calculate it -- the exact optimization of modularity is NP-hard, and the
current algorithm definitely has an exponential complexity so I wouldn't
even try it. The size of the RAM does not matter by the way, it is not the
RAM that is the bottleneck here but the number of possible partitions that
have to be evaluated. Also, the number of cores does not matter either
because the GLPK solver uses a single core only.
Best,
Tamas