Classification of olives from Moroccan regions by using direct FT-IR analysis: Application of support vector machines (SVM)

The aim of this work was to characterize and classify three close regions of olives by direct analysis on the olive without any preliminary treatment. This study was focused on the olive samples picked in the three zones: named Bazaza, oled ayad and oled hamdan, in the Moroccan region of Beni Mellal. All samples were also analysed by FT-IR spectroscopy, the spectral data were subjected to a preliminary derivative transform based on the gap segment algorithm to reduce the noise and extract a largest number of analytical information from spectra. A multivariate statistical procedure based on cluster analysis (CA) coupled to support vector machines (SVM), was elaborated, providing an effective classification method. On the basis of a hierarchical agglomerative CA and principal component analysis (PCA), three distinctive clusters were recognized. The SVM procedure was then applied to classify samples from the same regions. The model resulted able to separate the three classes and classify new objects into the appropriate defined classes with a percentage prediction of 93%. The results showed that FTIR spectroscopy coupled with chemiometric methods are an interesting technique for classifying olive samples according to their geographical origins.