CODEWITHSUNDEEP
rk-means
user3079364
user3079364

Reputation: 181

To find intersection of clusters in R

Let's assume I have done several operations and created cluster vectors of correlation values shown below

D <- matrix(rexp(10*10,rate=.1), ncol=10) #create a randomly filled 10x10 matrix
C <- matrix(rexp(10*10,rate=.1),ncol=10)

DCor <- cor(D) # generate correlation matrix
CCor <- cor(C)

DUpper<- DCor[upper.tri(DCor)] # extract upper triangle
CUpper<- CCor[upper.tri(CCor)]

ClusterD <- kmeans(DUpper,3) # cluster correlations
ClusterC <- kmeans(CUpper,3)

ClusterC <- cbind(c(1:45),matrix(ClusterC$cluster)) # add row numbers as column
ClusterD <- cbind(c(1:45),matrix(ClusterD$cluster))

I would like to generate a matrix shows the intersection of each cluster group. In this matrix, 5 rows belong to both C1 and D2 group.

enter image description here

How can I generate a matrix like this?

Upvotes: 0

Views: 487

Answers (1)

Sraffa
Sraffa

Reputation: 1668

Before the cbind lines, you could do:

table(ClusterC$cluster, ClusterD$cluster)

Upvotes: 1

Related Questions