CDPCA: 10 years after

Abstract

Clustering and Disjoint Principal Component Analysis (CDPCA) is a constrained principal component analysis for multivariate numerical data. The main goal is to detect clusters of objects and, simultaneously, to fi nd a partitioning of variables such that the between cluster deviance in the reduced space of such partition is maximized. The partition formed by a disjoint set of the original variables identifi es the groups of variables belonging to the CDPCA components. Recently, this methodology has been implemented in a R-function called CDpca. In this work, we review some theoretical issues of the CDPCA model and present two applications on real data sets using the R-function CDpca.