Simple Component Analysis Based on RV Coefficient

Among linear dimensional reduction techniques, Principal Component Analysis (PCA) presents many optimal properties. Unfortunately, in many applicative case PCA doesn't produce full interpretable results. For this reason, several authors proposed methods able to produce sub optimal components but easier to interpret like Simple Component Analysis (Rousson and Gasser, (2004)). Following Rousson and Gasser, in this paper we propose to modify the algorithm used for the Simple Component Analysis by introducing the RV coefficients (SCA-RV) in order to improve the interpretation of the results.