Discrimination of yerba mate (Ilex paraguayensis St. Hil.) samples according to their geographical origin by means of near infrared spectroscopy and multivariate analysis

Abstract

Near infrared reflectance (NIR) spectroscopy combined with multivariate data analysis was used to discriminate between the geographical origins of yerba mate (Ilex paraguayensis St. Hil.) samples. Samples were purchased from the local market and scanned in the NIR region (1100-2500 nm) in a monochromator instrument in reflectance. Principal component analysis (PCA), partial least squares discriminant analysis (PLS-DA) and linear discriminant analysis (LDA) were used to classify the samples based on their NIR spectra according to their geographical origin. Full cross validation was used as validation method when classification models were developed. The overall classification rates obtained were 76 and 100% using PLS-DA and LDA, respectively. The results demonstrated the usefulness of NIR spectra combined with multivariate data analysis as an objective and rapid method to classify yerba mate samples according to their geographical origin. Nevertheless, NIR spectroscopic might provide initial screening in the food chain and enable costly methods to be used more productively on suspect specimens.

Authentication of yerba mate according to the country of origin by using Fourier transform infrared (FTIR) associated with chemometrics

“…With respect to PLS-DA, Cozzolino et al (2010) analysed the entire NIR reflectance spectra of yerba mate and obtained results similar of those of the present study. As reported by Marcelo, Martins, Pozebon, Dressler, et al (2014), variable selection was necessary in order to improve the classification of yerba mate by PLS-DA and SVM-DA of the NIR reflectance spectra.…”

“…From an economic perspective, authenticity is essential to avoid unfair competition, because producers associate their brands with a given region in order to obtain market recognition, commercial advantages and better prices (Karoui & De Baerdemaeker 2007). Cozzolino et al (2010) identified yerba mate from Argentina and Uruguay by means of near-infrared spectroscopy (NIR) and multivariate analysis of the NIR spectra, whereas linear discriminant analysis (LDA) and partial least square discriminant analysis (PLS-DA) were used. Using a similar approach, Marcelo, Martins, Pozebon, Dressler, et al (2014) identified yerba mate from Argentina, Brazil, Uruguay and Paraguay.…”

“…Pre-processing using other normalisation methods, such as SNV or MSC, and the data mean centred only were also evaluated, but the results obtained were not satisfactory. In previous studies carried out using NIR, Cozzolino et al (2010) applied the second derivative to remove the constant baseline (offset) and the baseline slope. These authors used the SNV method to normalise the data.…”

Wavelength selection framework for classifying food and pharmaceutical samples into multiple classes

“…The first dataset refers to Yerba mate ( Ilex paraguariensis ), which has been cultivated and consumed in southern Latin America for centuries because of its medicinal properties . Previous works dealing with the problem of classifying yerba mate samples include Choi et al , who reported a successful discrimination between 11 Ilex species from different origins by means of multivariate analysis of their metabolite fingerprints obtained by 1H NMR, and Cozzolino et al , who applied PLS‐DA (Partial Least Squares – Discriminant Analysis) regression and LDA to discriminate between samples of yerba mate coming from three different countries. The application of the method proposed in this paper to 54 samples of yerba mate proceeding for four countries increased the classification accuracy average 36.75%, from 0.7156 to 0.9786, and retained average 3.54% of the original wavelengths.…”

Multivariate Analysis of the Profile of Elements Concentrations in the Yerba Mate (Ilex paraguariensis) Marketed in the MERCOSUR

“…Sendo assim, as distâncias são apresentadas comumente de forma relativa. 11 Para a erva mate, por exemplo, a HCA foi capaz de identificar o produto comercial por país de origem mediante o perfil de reflexão na região do infravermelho próximo (NIR) 12,13 e diferenciar as plantas da família Ilex através do perfil da composição das plantas obtido mediante espectrometria de ressonância magnética nuclear de hidrogênio (RMN de H 1 ). 14 A Análise por Componentes Principais (PCA) é a base fundamental da maioria dos métodos de tratamento multivariado de dados.…”