Use clustering consensus to infer cluster number.
CountClusters(data, tol = 0.01, range = 2:20, eigengap = TRUE, n_comp)
data | a symmetric nonnegative similarity matrix |
---|---|
tol | cutoff for lambda zero |
range | a vector specifying the min and max number of clusters to iteratively test when building the consensus matrix |
eigengap | whether or not to use the max eigengap (upper bound) cluster count |
n_comp | the number of similarity components to use for ensemble clustering |
the number of clusters and the vector of eigenvalues