D. Blanco-Gomis, I. Herrero-Sánchez and J. J. Mangas Alonso
Analytical techniques (HPLC and flow-injection analysis) for determining sugars, organic acids, polyphenols and pectins in apples, were employed along with chemometrics in the ripening and classification studies of cider apples. The use of principal component analysis allowed the authors to reduce the dimensionality of the data matrix; three new variables were obtained that accounted for 76% of variance. The projection of the apple cultivars in the reduced space allowed us to visualize the data structure on the basis of the degree of ripening and technological characteristics of the cider apple varieties monitored. Linear discriminant analysis computed a canonical variable with a prediction capacity of 93%, using three groups for cancellation in order to validate the method. The use of modelling techniques, such as SIMCA and partial least squares made an adequate grouping of apple cultivars feasible on the basis of their degree of ripening.