A convenient classification method of olive oils from different regions of Morocco without any chemical separation is proposed. A series of 29 virgin olive oils, all belonged to the variety known as “Picholine Marocaine” were sampled during the harvest season 2009, in the zones of Meknes (MEK), Ksiba (KS), Bradia (BR) and F.Bensalah (FBS). A preliminary treatment of the FTIR data was done by a derivative elaboration to reduce the noise and extract a largest number of analytical information from spectra. Cluster Analysis (CA) coupled to Principal Component Analysis (PCA) provides an effective grouping in four distinctive clusters. A PLS-DA procedure was then defined to classify samples from the same regions but picked in different times or unknown olive oil samples. The model was optimized by applying the Martens' Uncertainty Test which provided to select the wavelength zones carrying the most useful analytical information. Validation of the model gave a high values of correlation (R2 above 0.989) and low prediction errors (RMSEP below 0.049). The successful application of the final model proved the ability of the chemometric techniques applied on FTIR data for a rapid and accurate classification of olive oils without the need of sample preparation.
CLASSIFICATION OF MOROCCO OLIVE OILS BY FIRST DERIVATIVE FOURIER TRANSFORMED INFRARED SPECTROSCOPY
DE LUCA M;IOELE, Giuseppina;RAGNO G.
2010-01-01
Abstract
A convenient classification method of olive oils from different regions of Morocco without any chemical separation is proposed. A series of 29 virgin olive oils, all belonged to the variety known as “Picholine Marocaine” were sampled during the harvest season 2009, in the zones of Meknes (MEK), Ksiba (KS), Bradia (BR) and F.Bensalah (FBS). A preliminary treatment of the FTIR data was done by a derivative elaboration to reduce the noise and extract a largest number of analytical information from spectra. Cluster Analysis (CA) coupled to Principal Component Analysis (PCA) provides an effective grouping in four distinctive clusters. A PLS-DA procedure was then defined to classify samples from the same regions but picked in different times or unknown olive oil samples. The model was optimized by applying the Martens' Uncertainty Test which provided to select the wavelength zones carrying the most useful analytical information. Validation of the model gave a high values of correlation (R2 above 0.989) and low prediction errors (RMSEP below 0.049). The successful application of the final model proved the ability of the chemometric techniques applied on FTIR data for a rapid and accurate classification of olive oils without the need of sample preparation.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
