Graph clustering using the pairwise graphon distances and spectral clustering
graphonSpectralClustering(allAdj, nbClusters, sig = 0.1, nbCores = 1)
allAdj
: list of adjacency matricesnbClusters
: number of clusters to be foundsig
: parameter for Gaussian kernel used for the similarity matrixnbCores
: number of cores for parallelization.list with the obtained graph clusteirng ($clust) and the matrix with the pairwise graphon distances between all pairs of networks
theta <- list(pi=c(.5,.5), gamma=matrix((1:4)/8,2,2)) obs <- rCollectSBM(rep(10,4), theta)$listGraphs res <- graphonSpectralClustering(obs, 2, nbCores=1)
Useful links