These data are a simplified version of the segmentation data contained
in caret. There are three columns: A and B are predictors and the column
Class is a factor with levels "One" and "Two". There are three data sets:
one for training (n = 1009), validation (n = 300), and testing (n = 710).