The Gifi system of descriptive multivariate analysis

The Gifi system of analyzing categorical data through nonlinear varieties o f classical multivariate analysis techniques is reviewed. The system is cha racterized by the optimal scaling of categorical variables which is impleme nted through alternating least squares algorithms. The main technique of ho mogeneity analysis is presented, along with its extensions and generalizati ons leading to nonmetric principal components analysis and canonical correl ation analysis. Several examples are used to illustrate the methods. A brie f account of stability issues and areas of applications of the techniques i s also given.