Fourier transform infrared (FTIR) coupled to chemometrics was shown to be a useful method to classify and predict the quality of four commercial grade virgin olive oils (VOO). FTIR and physicochemical data were collected using a set of 70 samples representing extra virgin (EV), virgin (V), ordinary virgin (OV), and lampante (L) commercial grade olive oils collected in Beni Mellal region (central Morocco). Two partial least squares discriminant analysis (PLS-DA) models using physicochemical data and FTIR data were established and compared. The PLS-DA model using only physicochemical data was not accurate enough to distinguish satisfactorily among OV, V, and EV olive oil grades. On the contrary, the PLS-DA model on FTIR data was better in the calibration, able to describe 98 % of the spectral information and predicting 93 % of the VOO grades. In the external validation, this PLS-DA model accurately classified VOO commercial grades with prediction accuracy of 100 %. The proposed procedure is fast, nondestructive, simple, and easy to operate, and it is recommended for the quick monitoring of olive oil's quality.

Hirri A., De Luca M., Ioele G., Balouki A., Bassbasi M., Kzaiber F., Oussama A., Ragno G. (2015): Chemometric classification of citrus juices of Moroccan cultivars by infrared spectroscopy. Czech J. Food Sci., 33: 137-142.Fourier transform -infrared (FTIR) spectroscopy in connection with chemometric methodologies was used as a fast and direct analytical approach to classify citrus cultivars by the measurements on their juice. Modern multivariate analysis responds to the current needs for economic, simple, and fast methods able to classify new unknown samples with great accuracy. A set of 135 samples of citrus juice, representative of three cultivars (Hamlin, Muska, and Valencia), all picked in the same geographical area of Morocco, were analysed. Chemometric discrimination of the juice samples was achieved by principal component analysis (PCA) performed on the FTIR spectral data from the juice samples, showing an explained variance of 97.84% by considering only 2 PCs. A fully correct classification of the three Moroccan cultivars was then obtained by using Partial Least Square-Discriminant Analysis (PLS-DA) modelling procedure.

No abstract

In this study, the adulteration of Moroccan Picholine extra virgin olive oil with Arbequina virgin olive oil was monitored using the Fourier transform mid-infrared (FT-MIR) spectroscopy technique and chemometrics methodologies. To discriminate between olive oil that has been adulterated and unadulterated, principal component analysis (PCA) was utilized for qualitative analysis. We created the best calibration models for quantitative analysis using principal component regression (PCR) and partial least-squares regression (PLS). The first three principal components account for 95% of the overall variability, according to PCA analysis. PCA allows for the classification of the dataset into two groups: adulterated and unadulterated Moroccan Picholine olive oil. The application of the PLS and PCR calibration models for the quantification of adulteration demonstrates high-performance capabilities, as indicated by high values of correlation coefficients R2 greater than 0.999 and 0.995 and lower values of root mean square error (RMSE) less than 0.767 and 2.16 using PLS and PCR, respectively. According to our results, FT-MIR spectroscopy combined with chemometrics approaches can be used successfully as a simple, quick, and non-destructive method for the quantification and discrimination of adulterated olive oil.