find_clusters {cvxclustr} | R Documentation |
find_clusters
uses breadth-first search to identify the connected components of the corresponding
adjacency graph of the centroid differences vectors.
find_clusters(A)
A |
adjacency matrix |
## Clusterpaths for Mammal Dentition data(mammals) X <- as.matrix(mammals[,-1]) X <- t(scale(X,center=TRUE,scale=FALSE)) n <- ncol(X) ## Pick some weights and a sequence of regularization parameters. k <- 5 phi <- 0.5 w <- kernel_weights(X,phi) w <- knn_weights(w,k,n) gamma <- seq(0.0,43, length.out=100) ## Perform clustering nu <- AMA_step_size(w,n) sol <- cvxclust_path_ama(X,w,gamma,nu=nu) ## Construct adjacency matrix A <- create_adjacency(sol$V[[10]],w,n) find_clusters(A)