Premium
Classification of Slovak varietal white wines by volatile compounds
Author(s) -
Pet'ka Ján,
Mocák Ján,
Farkaš Pavel,
Balla Branko,
Kováč Milan
Publication year - 2001
Publication title -
journal of the science of food and agriculture
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.782
H-Index - 142
eISSN - 1097-0010
pISSN - 0022-5142
DOI - 10.1002/jsfa.979
Subject(s) - linear discriminant analysis , quadratic classifier , vintage , wine , mathematics , multivariate statistics , food science , chemistry , statistics , artificial intelligence , computer science , support vector machine , biochemistry
Selected volatile compounds in the most widely produced Slovak varietal white wines, Welschriesling, Grüner Veltliner and Müller Thurgau, were analysed over a 3 year period. The study aimed to find a combination of volatiles that could serve as an input for the classification of these wines using multivariate data analysis. Significant differences among the wine samples as regards variety, producer and vintage were initially examined using ANOVA. Two classification techniques were also employed, namely linear (LDA) and quadratic (QDA) discriminant analyses. LDA revealed five volatile compounds that enabled a satisfactory classification of the wines by variety, and 10 volatiles for the classification of producers. QDA was slightly less successful in distinguishing the varieties; however, the classification of producers was comparable with the LDA results when using a significantly smaller number of variables. © 2001 Society of Chemical Industry