Open AccessCharacterization and classification of wines according to geographical origin, vintage and specific variety based on elemental content: a new chemometric approach
Author(s) -
Ioana Feher,
Dana Alina Măgdaș,
Adriana Dehelean,
Costel Sârbu
Publication year - 2019
Publication title -
journal of food science and technology/journal of food science and technology
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.656
H-Index - 68
eISSN - 0975-8402
pISSN - 0022-1155
DOI - 10.1007/s13197-019-03991-4
Subject(s) - vintage , principal component analysis , linear discriminant analysis , wine , pattern recognition (psychology) , artificial intelligence , mathematics , fuzzy logic , hierarchical clustering , partition (number theory) , computer science , multivariate statistics , cluster analysis , statistics , data mining , chemistry , food science , biochemistry , combinatorics
A highly informative chemometric approach using elemental data to distinguish and classify wine samples according to different criteria was successfully developed. The robust chemometric methods, such fuzzy principal component analysis (FPCA), FPCA combined with linear discriminant analysis (LDA), namely FPCA-LDA and mainly fuzzy divisive hierarchical associative-clustering (FDHAC), including also classical methods (HCA, PCA and PCA-LDA) were efficaciously applied for characterization and classification of white wines according to the geographical origin, vintage or specific variety. The correct rate of classification applying LDA was 100% in all cases, but more compact groups have been obtained for FPCA scores. A similar separation of samples resulted also when the FDHAC was employed. In addition, FDHAC offers an excellent possibility to associate each fuzzy partition of wine samples to a fuzzy set of specific characteristics, finding in this way very specific elemental contents and fuzzy markers according to the degrees of membership (DOMs).