Two vectors of information are needed to produce the optimally scaled variable: the initial variable scaling vector x and the target vector y. For convenience, both vectors are first sorted on the ...