Merge States of an EMM
Merge several clusters/states of an EMM into a single cluster/state.
"merge_clusters"(x, to_merge, clustering = FALSE, new_center = NULL, copy=TRUE)
"EMM"object. Note that the function will change this EMM!
- vector of names of the states/clusters to merge. The
name of the first state in
to_mergeis used as the name for the new state representing the merged states.
to_mergea vector with cluster assignments as created by a clustering algorithm?
- supply new centers for the merged clusters.
New centroids are automatically
computed. If (pseudo) medoids are used, new medoids should be supplied.
If none is supplied, the medoid of the cluster in
to_mergewhich has the most assigned observations is used as the new medoid (warning: this is probably not a good medoid!)
- logical; make a copy of x before reclustering? Otherwise the function will change
Returns the changed EMM with the states/clusters merged invisibly.
copy=FALSEthen it returns a reference to the changes object passed as
data("EMMTraffic") emm <- EMM(measure="eJaccard", threshold=0.2) build(emm, EMMTraffic) states(emm) ## create a new emm with states 1-3 merged emm_m123 <- merge_clusters(emm, c("1", "2", "3")) states(emm_m123)
Looks like there are no examples yet.