Is there a way to change the index on kmeans()$cluster? - r

Question

I am using kmeans() to create groups based on a score. The goal is to assign star ratings, so that the individuals with the highest scores get four stars, and the individuals with the lowest scores get 1 star. I would like to create the star variable based on the kmeans()$cluster value. However, as it stands, kmeans()$cluster indexes the clusters, but the index does not correspond to the relative position of the group.
Is there a way to manually assign the cluster indexes, or to set the index to be assigned in a certain order? I'm hoping to have kmeans()$cluster=1 for the low score group, kmeans()$cluster=2 for second lowest, etc.
id <- 1:500
set.seed(12); score <- runif(500, 0, 1)
dat <- data.frame(id, score)
km = kmeans(dat$score, 4, nstart=10)
plot(dat$score,
col = c(km$cluster),
main="K-Means result with 4 clusters",
pch=20,
cex=0.8)
dat$star <- km$cluster
plot(dat$score,
dat$star,
main="Score v. cluster number")

Answer 1

Any of these will yield a new cluster assignment vector such that 1 refers to the cluster with the smallest center, 2 the next and so on. The first is expressed solely in terms of fitted(km) whereas the second is expressed in terms of km$centers and km$cluster and the last is expressed in terms of fitted(km) and km$center
fit <- fitted(km)
factor(fit, labels = 1:nlevels(factor(fit)))
rank(km$centers)[km$cluster])
match(fitted(km), sort(km$centers))

Answer 2

Yes. You can just use a small table of what you want the values to be and use the original cluster number to look them up. Here is an example.
set.seed(2017)
KM3 = kmeans(iris[,1:4], 3)
KM3$cluster
[1] 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2
[39] 2 2 2 2 2 2 2 2 2 2 2 2 3 3 1 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3
[77] 3 1 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 1 3 1 1 1 1 3 1 1 1 1 1 1 3
[115] 3 1 1 1 1 3 1 3 1 3 1 1 3 3 1 1 1 1 1 3 1 1 1 1 3 1 1 1 3 1 1 1 3 1 1 3
The clusters are in an awkward order. I want the low numbered points to be in cluster 1, the middle in cluster 2 and the high numbered points in cluster 3. So I want to change all of the 1's to 3, the 2's to 1 and the 3's to 2.
Relabel = c(3,1,2)
KM3$cluster = Relabel[KM3$cluster]
KM3$cluster
[1] 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
[39] 1 1 1 1 1 1 1 1 1 1 1 1 2 2 3 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2
[77] 2 3 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 3 2 3 3 3 3 2 3 3 3 3 3 3 2
[115] 2 3 3 3 3 2 3 2 3 2 3 3 2 2 3 3 3 3 3 2 3 3 3 3 2 3 3 3 2 3 3 3 2 3 3 2
Just a little extra detail. It says Relabel = c(3,1,2) because I want 1 to become 3, so the first location has a 3. I want 2 to become 1, so the second location has a 1. And i want 3 to become 2 so the third location has a 2.