Prediction of oil content in the mesocarp of fruit from the macauba palm using spectrometry

Abstract

In the present work a model was developed for the prediction of the oil content of the mesocarp of fruit from the macauba palm, using visible and near infrared spectrometry. Reference values were determined the by Soxhlet method. The model was calibrated using spectral data from the mesocarp of macauba fruit by partial least squares regression, considering nine latent variables. The results of the calibration series were consistent with those of the validation series, registering an The coefficient of determination between the reference method and the developed model, systematic error between the predicted values and the measured values and root mean square error, for calibration and validation with independent data, respectively, equal to 0.8223, -9.2-14 and 5.917 and 0.7760, 7.081 and -0.064. VIS-NIR spectroscopy is a viable tool in the evaluation of genotypes in breeding programs for the macauba palm.