摘要：A comprehensive data collection of authentic “specialty” or “gourmet” oils, namely cold-pressed industrial virgin oils, was performed. Eight different botanical species, i.e., Almond, Apricot, Avocado, Hazelnut, Mosqueta rose, Rosehip, Sunflower, and Walnut oils were studied plus Olive oil as the gold standard of cold-pressed virgin oils. Two different analytical approaches are proposed to rapidly verify the botanical species of the oil-based raw material. The first approach is based on a multivariate statistical analysis of conventional analytical data, namely their fatty acid composition. These data have been re-elaborated in a multivariate way by Principal Component Analysis (PCA) and classification methods. The second approach proposes a fast and non-destructive spectrophotometric analysis to determine the color of these oils to discriminate among different species. In this regard, the raw diffuse reflectance spectra (380–780 nm) obtained by a UV-Vis spectrophotometer with an integrating sphere was considered and elaborated by chemometrics. This information was compared with the results obtained by the most common approach based on the CIELab parameters. A data fusion of chromatographic and spectral data was also investigated. Either fatty acid composition or color of these oils demonstrated to be two promising markers of their botanical authenticity.
关键词：specialty oils; fatty acids; untargeted spectroscopic fingerprint; principal component analysis; CIELab parameters; data fusion specialty oils ; fatty acids ; untargeted spectroscopic fingerprint ; principal component analysis ; CIELab parameters ; data fusion
