TraMineRextras (version 0.6.8)

pamward: PAM from k-solution of hierarchical clustering

Description

Runs a pam clustering (pam) from the solution in k groups of a hierarchical clustering (agnes) .

Usage

pamward(diss, k=3, method="ward", dist)

Value

An object of class "pam". See pam.object for details.

Author

Gilbert Ritschard

Arguments

diss

Distance matrix or object.

k

Integer. Number of clusters.

method

Method for the hierarchical clustering (see agnes).

dist

Deprecated. Use diss instead.

Details

The function first runs the hierarchical clustering, retrieves the medoids of the solution for the provided k and uses those medoids as start centers for the pam partitioning.

See Also

agnes and pam.

Examples

library(cluster)
data(actcal)
actcal.seq <- seqdef(actcal[1:200,13:24])
actcal.ham <- seqdist(actcal.seq, method = "HAM")
clust <- pamward(actcal.ham, k = 4)
table(clust$clustering)