Takes a subtree of a dendrogram object such as generated by hclust according to a prespecified number of clusters or a prespecified height.
subtree(tree, k = NULL, h = NULL)
The tree object which contains the attributes merge
and height
in the same way as an hclust object.
The number of clusters at which to cut.
The height at which to cut.
An hclust object.
Does the same as cutree with two differences: Firstly it gives back an entire tree object, i.e. an object with attributes merge
, height
, labels
and order
, as well as data
, which contains the cluster ids. Secondly the cluster ids are chosen by the heights at which the clusters were built.
# NOT RUN {
hc <- hclust(dist(USArrests), "ave")
hcs <- subtree(hc, k = 7)
hcs2 <- subtree(hc, h= 30)
attributes(hcs)
fluctile(table(hcs$data, cutree(hc, k=7)))
par(mfrow=c(1,3))
plot(hc)
plot(hcs)
plot(hcs2)
# }
Run the code above in your browser using DataLab