Example

data(iris)
library(ggplot2)
library(caret)

## Loading required package: lattice

inTrain<- createDataPartition(y=iris$Species,p=0.7,list=FALSE)
training<- iris[inTrain,]
testing<- iris[-inTrain,]

Cluster with k-means

kMeans1<- kmeans(subset(training,select=-c(Species)),center=3)
training$clusters<- as.factor(kMeans1$cluster)
qplot(Petal.Width,Petal.Length,col=clusters,data=training)

#Compare to real labels

table(kMeans1$cluster,training$Species)

##    
##     setosa versicolor virginica
##   1      0         34         8
##   2     35          0         0
##   3      0          1        27

Build Predictor

modFit<- train(clusters~.,data=subset(training,select=-c(Species)),method="rpart")

## Loading required package: rpart

table(predict(modFit,training),training$Species)

##    
##     setosa versicolor virginica
##   1      0         35         9
##   2     35          0         0
##   3      0          0        26

#On the training set, there is diversity on prediction of virginica class

Apply on test

testClusterPred<- predict(modFit,testing)
table(testClusterPred,testing$Species)

##                
## testClusterPred setosa versicolor virginica
##               1      0         15         7
##               2     15          0         0
##               3      0          0         8

#Simiarly, the prediction on virginica does not perform well, but the other two class are good