R FactoMineR: Return cluster membership

Question

I have produced an HCPC object and would like to return a list of the observations and which clusters they have fallen into. I can't find a command to do this; does anyone know?

FYI, here is my work:

file <- read.csv("file", header=TRUE) 
library(FactoMineR) 
res.mca = MCA(file, graph=FALSE) 
# manually cut tree according to inertia gain 
res.hcpc = HCPC(res.mca, nb.clust=0)

Edit: A reproducible example:

library(FactoMineR)
data(tea)
res.mca  <- MCA(tea, quanti.sup = 19, quali.sup = 20:36)
res.hcpc <- HCPC(res.mca, nb.clust = 5)

It'd be nice to have `"file"` to work with (even more would be a small example showing the result you expect). — Arun, Feb 05 '13 at 17:16
Sure, thanks. This example should work - I'm mainly just trying to return the cluster membership detail from any given HCPC object: — Jeff Coughlin, Feb 05 '13 at 17:21
library(FactoMineR) data(tea) res.mca=MCA(tea,quanti.sup=19,quali.sup=20:36) res.hcpc=HCPC(res.mca, nb.clust=5) — Jeff Coughlin, Feb 05 '13 at 17:22
You should be able to edit your post and your comments. I've edited the post for you. — Arun, Feb 05 '13 at 17:25

score 4 · Answer 1 · edited Aug 01 '15 at 17:16

The output $data.clust gives the a dataframe with the input data and a column (the last one) with the cluster the individuals belong.

library(FactoMineR)
data(tea)
res.mca  <- MCA(tea, quanti.sup = 19, quali.sup = 20:36)
res.hcpc <- HCPC(res.mca, nb.clust = 5)
res.hcpc$data.clust

Note that if you want that the individuals are in the same order as in the input dataset, you should use the argument order=FALSE in HCPC :

res.hcpc <- HCPC(res.mca, nb.clust = 5, order=FALSE)
res.hcpc$data.clust

score 1 · Answer 2 · edited May 23 '17 at 12:33

Although this post was written 3 years ago, my answer can be of help for readers. I was just looking for a solution to the same question as I used HCPC for outliers detection and here is the idea I got: I extracted a subset from the dataframe $data.clust issued from hcpc(), based on the field "cluster". outliers<-subset(DHCPC$data.clust, clust==4)

I found the answer here: how to extract a subset of a data frame based on a condition involving a field?

R FactoMineR: Return cluster membership

2 Answers2