Procrustes analysis is a technique of comparing the alternative geometrical representations of a group of multivariate data or of the proximity matrix, for instance, two competing multidimensional scaling solutions for the latter. The two solutions are compared to each other using a residual sum of the squares criterion, which is minimized by permitting the coordinates corresponding to one solution to be rotated, reflected and translated relative to other. Generalized Procrustes analysis permits comparison of more than two alternative solutions by at the same time translating, rotating and reflecting them so as to optimize the predefined goodness-of-fit measure.