graphMomentsClustering function

Graph clustering method using graph moments

Graph clustering method based on graph moments by Mukherjee et al. (2017)

graphMomentsClustering(Networks, nbMoments = 3, nbClusters)

Arguments

• Networks: list of adjacency matrices
• nbMoments: order of the largest graph moments to be considered
• nbClusters: desired number of clusters

Returns

vector with the clustering of the networks

Examples

param <- vector('list', 3)
param[[1]] <- list(prop = 1/3, # component 1 :  alpha > beta
                   alpha = .04,
                   beta = .02,
                   deltaIn = 100,
                   deltaOut = 100,
                   R = 500
)
param[[2]] <- list(prop = 1/3, # component 2 : just permute alpha and beta ;
                   alpha = .01,
                   beta = .02,
                   deltaIn = 100,
                   deltaOut = .1,
                   R = 1000
)
param[[3]] <- list(prop = 1/3, # component 3 : alpha=beta
                   alpha = .015,
                   beta = .015,
                   deltaIn = .1,
                   deltaOut = .1,
                   R = 1000
)
obs <- sampleDPAMixture(M=20, param)
res <- graphMomentsClustering(obs$listAdj, 3, 3)
table(res, obs$graphGroups)