Shukai,
layout.drl supports edge weights, so you could try that. An
alternative is doing MDS, see ?cmdscale and maybe help.search("MDS").
But both these methods are approximate, obviously, most often you
cannot embed an n-dimensional (n>>2) graph into the 2-dimensional
plane and keep all the "dis
Thanks Gabor's fast reply.
In my research, every node has it's own vector of scores. So I can compute
correlation between every pair of nodes. I used the width and color of the
edge for this purpose. But visualizing the correlations by distance may be
clearer.
Best,
Shukai
Gábor Csárdi-2 wro
Shukai,
the force based layout algorithms (layout.drl,
layout.fruchterman.reingold, layout.graphopt, layout.kamada.kawai) are
likely to do this; although they are not explicitly required to place
hubs in the center, usually they do.
I am not sure what is the "correlation between two nodes". You m
Dear R users,
I am trying to draw a network using igraph package. I intend to place the
hub nodes (the ones with the relatively more connection with other nodes) in
the center of the graph. Also, the graph need to be in the fashion that the
higher the correlation between two nodes is , the clos
4 matches
Mail list logo