Re: [igraph] Fastgreedy Algorithm for Community Detection in igraph for RStudio - fatal error.

[Tamás Nepusz]

Hi,

> Is it possible, that I have a newer version of igraph or RStudio, which is 
> not compatible with the fastgreedy alg.? What can I do?
I strongly suspect that the problem lies in RStudio itself because all the 
community detection algorithms work perfectly fine for me if I use them from 
the command line version of R. But just to make sure, please send me the small 
network that you managed to crash RStudio with and I'll check on my machine.

> Then I have a second question: I’m also using the Optimal Modularity 
> Algorithm. But it needs a lot of time to compute (I started it 24 hours ago 
> and it’s still computing). Is there a possibility to make it faster?
No, the optimal modularity algorithm is pretty much unusable for larger graphs 
(more than 40-50 nodes I guess). This is because it uses integer programming in 
the background and solving a general integer programming problem has 
exponential time complexity. On the other hand, if the algorithm finishes, it 
will give you the global optimum (unlike other community detection algorithms 
which usually do not have any performance guarantees).

-- 
T.