®®®® SIIA Público

Título del libro: Quality Control In The Beverage Industry: Volume 17: The Science Of Beverages
Título del capítulo: Multivariate classification techniques to authenticate mexican commercial spirits

Autores UNAM:
MARIA GUADALUPE PEREZ CABALLERO; ALMA LUISA REVILLA VAZQUEZ;
Autores externos:

Idioma:
Inglés
Año de publicación:
2019
Palabras clave:

Authentication; Chemometrics; Mezcal; Multivariate classification; Pattern recognition; Tequila


Resumen:

A major challenge in the fight against branded-beverages counterfeiting is the development of analytical methodologies capable of discriminating between the types of a commercial product and, also, between correctly and incorrectly labeled products. Such methodologies require hyphenating knowledge from two scientific fields: analytical chemistry and computation science. This chapter reports on the use of a suite of statistical and computational tools (termed ?chemometrics? in the chemical field) to analyze and characterize commercial beverages. Their fundamentals are reviewed conceptually, along with some graphical examples to show typical classification outputs for each method under discussion. Some of their advantages and disadvantages are discussed, as well. Most examples here deal with tequila, which is one of the most popular alcoholic beverages in the world and whose production has increased substantially in recent years. Out of the 229 million liters produced in 2015, ca. 183 million ones were exported to over 120 countries worldwide. Unfortunately, counterfeiting is raising as well and, therefore, the development of advanced tools to assess the correct quality of a dubious batch is of paramount importance. Before studying the data treatment tools (which involve different levels of complexity; namely: pattern recognition, k-nearest neighbor, potential functions, discriminant analysis, soft-independent modeling of class analogy, partial least squares-discriminant analysis, classification and regression trees, random forest, artificial neural networks, and support vector machines) the most remarkable analytical methodologies employed so far for authenticating Mexican spirits are reviewed. © 2019 Elsevier Inc. All rights reserved.


Entidades citadas de la UNAM: