How do jmp choose the most optimal number of clusters, when you choose the option "number of clusters"? (hierarchical - ward)
Thankyou!
But the distance in our agglomeration schedule isn't increasing. What's the reason for that?