Estimate PAM clustering solution and optimal k using fpc::pamk
u.PAMK(x, krange = 2:10, criterion = "asw", usepam = ifelse(nrow(x) <
2000, TRUE, FALSE), scaling = TRUE, diss = inherits(data, "dist"),
metric = "euclidean", do.swap = TRUE, trace = 0, verbose = TRUE,
...)Input matrix / data.frame
Integer vector: Range of k values to try
String: Criterion to use for selecting k: "asw", "multiasw" or "ch". See fpc::pamk
Logical: If TRUE, use cluster::pam, otherwise use cluster::clara.
Default = TRUE
Logical or Numeric vector: If TRUE, scale input. If numeric vector of length equal to number of features, the features are divided by the corresponding value. Default = TRUE
Logical: If TRUE, treat x as a dissimilarity matrix, otherwise as a matrix of
cases by features. Default = TRUE, if x inherits from class dist, FALSE otherwise.
String: Dissimilarity metric to be used. Options: 'euclidean', 'manhattan'
Logical: If TRUE, perform the swap phase (See cluster::pam), as in the
original PAM algorithm. This is computationally intensive and can be skipped. Default = TRUE
Integer [0, 3]: Trace level for fpc::pamk
Logical: If TRUE, print messages to screen
Additional parameters to be passed to fpc::pamk and/or cluster::pam
rtClust object
Other Clustering: u.CMEANS,
u.EMC, u.H2OKMEANS,
u.HARDCL, u.HOPACH,
u.KMEANS, u.NGAS,
u.PAM, u.SPEC