Which statement is CORRECT about clustering methods?

A$.$ We can define and measure clustering accuracy on a unlabeled testing dataset.

B$.$ Clustering algorithms construct clusters where between$-$cluster variation $($BCV$)$

is small compared to within$-$cluster variation $($WCV$).$

C$.$ Data mining projects rarely use clustering as a preliminary step.

D$.$ In a clustering output, each data record belongs to one and only one of the

clusters.