Fourier transform infrared spectroscopy has been applied in tandem with multivariate statistical approaches, with the prospect of developing a methodology for the prediction of interesting traits in distinct clones of “Olea europaea L.” For this purpose, the infrared spectra of either the drupes or the corresponding olive oils of 6 distinct clones of the Cobrançosa cultivar, in 2 ripening stages (semi‐ripe and ripe), have been registered resorting to an attenuated total reflection accessory. These clones were previously evaluated with respect to antioxidant activity and olive oil yield, while multivariate analyses, namely, partial least squares regression and discriminant analysis have been applied to the spectral data collected to obtain a prediction model for the assessment of these traits in distinct clones through the infrared spectra.
This methodology leads to minimal erroneous classification rates in the validation procedure of 3.9% and 6.3% for olives and olive oils, respectively, using the first derivative of the spectra. Thus, the suitability of this methodology for the assessment of different clones has been shown with visible benefits for a time consuming and costly process such as clonal selection. Furthermore, this procedure might be extended for other cultivars or different species once proper calibration models are attained for each matrix.