Iteratively maximize bipartite modularity.

This function is based on the bipartite modularity as defined in "Modularity and community detection in bipartite networks" by Michael J. Barber, Phys. Rev. E 76, 066102 (2007) This function uses a slightly different implementation from the paper. It does not use the "adaptive BRIM" method for identifying the number of modules. Rather, it simply continues to iterate until the difference in modularity between iterations is less that 10^-4. Starting from a random initial condition, this could take some time. Use condorCluster for quicker runtimes and likely better clustering, it initializes the blue node memberships by projecting the blue nodes into a unipartite "blue" network and then identify communities in that network using a standard unipartite community detection algorithm run on the projected network. See condorCluster for more details that.

Usage

condorModularityMax(
  condor.object,
  T0 = cbind(seq_len(q), rep(1, q)),
  weights = 1,
  deltaQmin = "default"
)

Arguments

condor.object

is a list created by createCondorObject. condor.object$edges must contain the edges in the giant connected component of a bipartite network

T0

is a two column data.frame with the initial community assignment for each "blue" node, assuming there are more reds than blues, though this is not strictly necessary. The first column contains the node name, the second column the community assignment.

weights

edgeweights for each edge in edgelist.

deltaQmin

convergence parameter determining the minimum required increase in the modularity for each iteration. Default is min(10^-4,1/(number of edges)), with number of edges determined by nrow(condor.object$edges). User can set this parameter by passing a numeric value to deltaQmin.

Value

Qcoms data.frame with modularity of each community.

modularity modularity value after each iteration.

red.memb community membership of the red nodes

blue.memb community membership of the blue.nodes

Examples

r = c(1,1,1,2,2,2,3,3,3,4,4);
b = c(1,2,3,1,2,4,2,3,4,3,4);
reds <- c("Alice","Sue","Janine","Mary")
blues <- c("Bob","John","Ed","Hank")
elist <- data.frame(red=reds[r],blue=blues[b])
condor.object <- createCondorObject(elist)
#randomly assign blues to their own community
T0 <- data.frame(nodes=blues,coms=1)
condor.object <- condorModularityMax(condor.object,T0=T0)