Characterization of Catalan red wines by pattern recognition methods

Authors

  • M. S. Larrechi
  • J. Guasch
  • F. X. Rius
  • A. Solé

DOI:

https://doi.org/10.5073/vitis.1987.26.172-182

Keywords:

red wine, differentiation, analysis, manganese, magnesium, iron, sodium, ethano, acidity, statistics, Spain

Abstract

52 red wines produced in the Tarragona province (Catalonia, Spain), 1983 vintage, belonging to 4 different regions, namely Priorat, Terra Alta, Camp de Tarragona and Falset, have been characterized and differentiated according to their geographic origin on the basis of 17 parameters measured for each sample and 3 pattern recognition methods: Statistical Quadratic Discriminant Analysis, K-Nearest Neighbour and Linear Learning Machine. The meta! ions manganese, magnesium, iron and sodium, along with the enological parameters ethanol content, alkalinity of ashes and titratable acidity have been found as the most relevant ones in this study. The lack of varietal differences together with quite similar climatic and cultural conditions due to the geographic contiguity of the studied zones leads to a maximum prediction ability of 87 % with the Linear Learning Machine method.

Downloads

Published

2015-12-16

Issue

Section

Article