CODEWITHSUNDEEP
rigraphbipartite
Anum Fatima
Anum Fatima

Reputation: 1

Sampling a bipartite graph on given set of vertices

I want to sample a bipartite graph on a given set of vertices rather than just giving the number of vertices of each type like in sample_bipartite(n1, n2, p). To be more precise, I want to simulate a graph with two groups of vertices and create a graph by sampling vertices within each group and across those groups of vertices. My current code is:

gw1 = sample_gnp(n1, p1, directed = FALSE, loops = FALSE)
n_e11=gsize(gw1)

E(gw1)$group <- rep(1,n_e11)

gw2 = sample_gnp(n2, p2, directed = FALSE, loops = FALSE)
n_e22=gsize(gw2)

E(gw2)$group <- rep(2,n_e22)

Next, I want to sample a bipartite graph between the vertices of gw1 and gw2. Currently, I have written the following code to achieve this goal.

gb12 = sample_bipartite(g_n[1], g_n[2], p = p[1,2])
n_e12=gsize(gb12)

E(gb12)$group <- rep(12,n_e12)
  
gw12 = gw1 + gw2
gg12 <- gb12 %u% gw12
  
E(gg12)$group_1[is.na(E(gg12)$group_1)] <- 0
E(gg12)$group_2[is.na(E(gg12)$group_2)] <- 0
E(gg12)$group <- E(gg12)$group_1 + E(gg12)$group_2

This gives me a graph I desire somehow, but I am not sure I am doing the right thing technically. If I could use the vertex set from the first two graphs to sample a bipartite graph, it would be technically what I want. I would appreciate if I can get some guidance on this.

Upvotes: 0

Views: 49

Answers (1)

Szabolcs
Szabolcs

Reputation: 25703

I want to simulate a graph with two groups of vertices and create a graph by sampling vertices within each group and across those groups of vertices

A graph that contains connections within the two groups as well is not bipartite, by definition.

You seem to be looking for sample_sbm(), which is like sample_gnp(), but we can specify a different p for each pair of groups, as well as within each group. You can have any number of groups.

Recommended reading: https://en.wikipedia.org/wiki/Stochastic_block_model

Upvotes: 0

Related Questions