Classification of moroccan olive cultivars by linear discriminant analysis applied to ATR-FTIR spectra of endocarps

• Published in: International journal of food science & technology Vol. 47; no. 6; pp. 1286 - 1292
• Main Authors: De Luca, Michele, Terouzi, Wafa, Kzaiber, Fouzia, Ioele, Giuseppina, Oussama, Abdelkhalek, Ragno, Gaetano
• Format: Journal Article
• Language: English
• Published: Oxford, UK Blackwell Publishing Ltd 01.06.2012; Wiley-Blackwell; Oxford University Press
• Subjects: algorithms; Biological and medical sciences; Chemometrics; Classification; Cultivars; Discriminant analysis; endocarp; Food industries; Food science; Fourier transform infrared spectroscopy; Fundamental and applied biological sciences. Psychology; linear discriminant analysis; olives; Principal components analysis; Reflectance; spectral analysis; variance; Discriminant analysis; Fourier transformation; Oleaceae; endocarp; Olea europaea; Olive; Infrared spectrometry; olives; Dicotyledones; Angiospermae; Classification; Spermatophyta; fourier transform infrared spectroscopy; linear discriminant analysis; Chemometrics
• ISSN: 0950-5423; 1365-2621
• DOI: 10.1111/j.1365-2621.2012.02972.x

Summary The potential of FTIR combined with chemometrics was studied to classify five Moroccan varieties of olives by analysis on the endocarps. Attenuated total reflectance (ATR) enabled the samples to be examined directly in the solid state. The spectral data were subjected to a preliminary derivative elaboration based on the Norris gap algorithm to reduce the noise and extract larger analytical information. Linear discriminant analysis (LDA) was adopted as classification method, and Principle component analysis (PCA) was employed to compress the original data set into a reduced new set of variables before LDA. The calibration set was built by using the IR data from seventy‐five samples scanned in reflectance mode, and the ranges 3000–2400 and 2300–600 cm−1 were selected because furnishing the most useful analytical information. PCA allowed clustering the samples in five classes by using the first two principal components with an explained variance of 98.16%. Application of LDA on an external test set of twenty‐five samples enabled to classify them into five variety groups with a correct classification of 92.0%.