Market Segmentation by Dual Scaling Through Generalized Forced Classification

Abstract

Forced classification, a technique for discriminant analysis of categorical data by dual scaling, is first presented together with its seven mathematical properties and its generalizations to dual scaling of modified data matrices. Simple and generalized forms of forced classification are then applied to hypothetical cases of market segmentation, namely, one-way classification, bipolar classification, conditional classification, classification by expert’s knowledge and multi-way classification. These examples suggest also other potential applications of dual scaling to market segmentation research.