I am using the clusterboot() function from the fpc R package to access cluster stability.

clusterboot(data_scaled, B=20, clustermethod=kmeansCBI, krange=2:10)

The krange was specified as 2:10

In the results there is the output “Number of clusters found in data: 4”

I am assuming that the function calculates the optimal number of clusters. However, I don`t know by which method (silhouette? Calinski-Harabasz Index?)

Can anyone tell me?

Thanks in advance