Il confronto dei vincoli ultrametrici di criteri classificatori mediante l’Analisi in Componenti principali rispetto ad un sottospazio di riferimento - A CPCA-based Comparison of Ultrametric Constraints from Clustering Criteria

We consider an Euclidean distance matrix with an external system of explanatory variables. The Constrained Principal Component Analysis on the Euclidean distance matrix and on the ultrametric distance matrix, on the same units, leads to single out the amount of the ultrametric constraint on each pair of units. We suggest a strategy to point out the cluster criterion having the smallest distortion due to the ultrametric constraints.