A method of scaling for a mixed set of discrete and continuous variables

Abstract

A method of equalized scaling is proposed for taxonomic or classificatory systems, designed so that problems involving a mixed set of variables (some quantitative and others multistate qualitative) can be handled automatically and so that variables of neither class will unduly dominate the classifying process. The method is formulated in terms of Euclidean distance in n-dimensional space but is easily adapted to a system using similarity coefficients. Other advantages of the method are discussed. © Oxford University Press.