Application of fourier transform mid-infra-red attenuated total reflectance (ft-mir-atr) for the authentication of Maltese extra virgin olive oil

Abstract

The price of extra virgin olive oil, a universally used natural product, depends on its botanical source and its production environment, causing extra virgin olive oil to be vulnerable for adulteration through mislabelling and inappropriate fraudulent production. The application of FT-MIR-ATR spectra in conjunction with several chemometric methods was found to provide a cheap, fast, and reliable way for the discrimination of Maltese EVOOs from non-Maltese EVOOs. Due to the high level of similarity and collinearity, the application of unsupervised PCA models was deemed to be unsatisfactory when it comes to discrimination of geographical origin. Application of supervised methods of classification namely PLSDA, ANN, LDA and SVM, showed to be highly effective in classifying and discriminating local and non-local EVOOs samples. The use of variable selection methods significantly increased the effectiveness of PLS-DA models when compared to no variable selection. ANN, SVM and LDA models were also shown to offer similar classification rates to PLS-DA models, giving further confidence in the application of FT-MIR.