On this note, I've wondered about whether there's a cheap way to optimize this.
In these projections there is usually a complete graph between all co-attendees of an event (or co-members at l1 of whatever at l2). But these are rarely meaningful on their own. Thus, whenever I do a one mode projection, the first thing I typically do is filter to edge weight >=2. Now, doing this cuts out about 80-90% or more of the ties in one fell swoop. Would it not be possible to have such a technique embedded at the lowest levels so that I am only returned networks where there is >1 weight to the edge? It might require a different algorithm to optimize, but it would certain mean significantly less edges in the resulting graphs. Take care, BERNiE Dr Bernie Hogan Research Fellow, Oxford Internet Institute University of Oxford http://www.oii.ox.ac.uk/people/hogan/ On 2013-08-13, at 2:21 PM, Gábor Csárdi <[email protected]> wrote: > Hi, > > unfortunately getting only one projection is not supported. It was already > requested before: > https://github.com/igraph/igraph/issues/307 > > I will code it up at some point, but probably not very soon. You can try > modifying the C code of igraph, or just write the function for yourself, it > is probably not very hard, and you could use graph.neighborhood to get the > second neighbors of the type-1 nodes. > > Gabor > > On Tue, Aug 13, 2013 at 8:58 AM, MIKOLAJCZAK, Alexis > <[email protected]> wrote: > > Hi, > > I'm also experiencing memory limitations on my pc for the projection of a > bipartite network. > > Is it possible to choose only one projection (in this case, the first one, > hopefully the one I want to look at) ? > > > bipartite.projection.size(g.bip, types = NULL) > > > > $vcount1 > [1] 2492 > > $ecount1 > [1] 803734 > > $vcount2 > [1] 46857 > > $ecount2 > [1] 377291985 > > Thank you for your answer. > Sincerly > > > -- > Alexis MIKOLAJCZAK > > Tél : 04.79.26.65.72 > > Conservatoire botanique national alpin > > Siège : Domaine de Charance – 05000 Gap – Tél. 00 33 (0)4 92 53 56 82 – Fax. > 00 33 (0)4 92 51 94 58 > > Antenne Alpes du nord / Ain : 148 rue Pasteur –73000 Chambéry – Tél. 00 00 33 > (0)4 79 33 45 04 > > Site internet: http://cbn-alpin.fr/– Mail : [email protected] > > SINP PACA : http://flore.silene.eu - SINP Rhône-Alpes : http://www.pifh.fr/ > > > > > -- > Alexis MIKOLAJCZAK > > Tél : 04.79.26.65.72 > > Conservatoire botanique national alpin > > Siège : Domaine de Charance – 05000 Gap – Tél. 00 33 (0)4 92 53 56 82 – Fax. > 00 33 (0)4 92 51 94 58 > > Antenne Alpes du nord / Ain : 148 rue Pasteur –73000 Chambéry – Tél. 00 00 33 > (0)4 79 33 45 04 > > Site internet: http://cbn-alpin.fr/– Mail : [email protected] > > SINP PACA : http://flore.silene.eu - SINP Rhône-Alpes : http://www.pifh.fr/ > > > _______________________________________________ > igraph-help mailing list > [email protected] > https://lists.nongnu.org/mailman/listinfo/igraph-help > > > _______________________________________________ > igraph-help mailing list > [email protected] > https://lists.nongnu.org/mailman/listinfo/igraph-help
_______________________________________________ igraph-help mailing list [email protected] https://lists.nongnu.org/mailman/listinfo/igraph-help
