Authentication of tequilas using pattern recognition and supervised classification

Abstract

[Abstract] Sales of reputed, Mexican tequila grown substantially in last years and, therefore, counterfeiting is increasing steadily. Hence, methodologies intended to characterize and authenticate commercial beverages are a real need. They require a combination of analytical characterization and chemometric tools. This work reports concisely on the former and focus on the chemometric tools employed so far in connection with them. Further, a practical case study presents the classification capabilities of nine supervised classification methods to differentiate white, rested, aged and extra-aged tequilas. The largest set of certified tequilas employed so far was considered. In general, non linear methods performed best than linear ones (accuracy higher than 94% in both training and validation). The case study demonstrates that it is possible to develop fast, cheap, easy to implement and reliable analytical methodologies to authenticate and classify samples of tequilas.