“…In the recent years, statistical learning models like Support Vector Machines (SVM) and Linear Discriminant Analysis (LDA) have played an important role for characterizing differences between a reference group of patterns and the population under investigation [1,4,13,24,21,22,23,12]. In general, the basic pipeline to follow in this subject is: (a) Dimensionality reduction; (b) Choose a learning method to compute a separating hypersurface, that is, to solve the classification problem; (c) Reconstruction problem, that means, to consider how good a low dimensional representation might look like.…”