Discriminant multiple correspondence analysis

The proposed approach leads to analyzing the associations between a set of quantitative variables and several qualitative variables measured on a same set of individuals. In a decision-making context, the proposed method can be considered as a generalization of discriminant analysis to the multiple groups’ variables case. It’s described as a principal component analysis of the centres of gravity tables. The decomposition of its duality diagram on a double diagram illustrates the link and the passing properties between multiple correspondence analysis and discriminant analysis. An application resulting from real data illustrates the utility of the discrimination model thus defined.