ABSTRACT
The comparison of partial points related to a given group poses no problem as the partial transition relation is applied in exactly the same way as the transition relation in Principal component analysis. Much multiple factor analysis use involves comparing subtables, each made up of a group of variables. It becomes difficult to analyse visually the superimposed representation of partial clouds when there are many individuals and/or groups. However, there is no property which makes it possible directly to compare partial points related to different groups inasmuch as they result from weighted means of different points. A series of indicators can be obtained by decomposing the inertia of the cloud bringing together all of the partial points. The coordinates of the partial points can therefore be calculated from the variables’ coordinates. Interpretations are mostly based on partial transition relations. They can be used to compare the partial points related to a given group directly.
