| Abstract: | This study deals with the development of a method for classification of yerba mate (Ilex paraguariensis) using attenuated total-reflectance Fourier transform infrared (ATR-FTIR) and multivariate analysis. Fifty-four brands of yerba mate from southern South America were analysed in order to classify the commercialised yerba mate according to the respective country of yerba mate processing. The yerba mate was ground in a cryogenic mill, and the reflectance was directly measured in the region ranging from 4000 to 650 cm?1. Different pre-processing algorithms and three methods of multivariate analysis were investigated, including principal component analysis (PCA), partial least square discriminant analysis (PLS-DA) and support vector machine discriminant analysis (SVM-DA). The yerba mate classification was 100% correct when the reflectance spectra were pre-treated (derived at first order, normalised by standard normal variation, smoothed and mean centred) and analysed using the SVM-DA method. |
