Multivariable analytical method based on NIR spectroscopy and chemometric tools to differentiate between '100% agave' and 'mixed' white tequila

Abstract

Tequila, one of Mexico's most representative spirits, is threatened by adulteration and counterfeiting practices that affect both consumers and the industry. The distinction between the categories of '100% agave' and 'mixed' white Tequila is crucial to guarantee its quality, cultural and commercial value. In this study, a multivariate analytical method based on near infrared spectroscopy (NIR) and chemometric tools is proposed to address this problem, using advanced mathematical models, including principal component analysis (PCA), partial least squares regression (PLSR), k-nearest neighbors (kNN), partial least squares regression discriminant analysis (PLS-DA), soft independent modeling of class analogy (SIMCA) and support vector machine learning (SVM). Results show that the SVM and PLS-DA models achieved perfect classifications (sensitivity, specificity and precision = 1), evidencing their effectiveness in the authentication of Tequila categories. Additionally, the SVMR regression model demonstrated outstanding performance in predicting alcohol content, with a coefficient of determination (R²) of 1.0. This approach offers an efficient, fast and non-destructive alternative to traditional quality control methods such as chromatography. Its implementation in the tequila industry and routine laboratories can significantly improve fraud detection and guarantee product authenticity. This study highlights the potential of NIR spectroscopy and chemometrics as key tools to strengthen quality control in the spirits industry.