PREDICTION OF VIRGIN OLIVE OIL SENSORY PR OFILE

The aim of this work is to present a statistical procedure to predict the extra virgin olive oil sensory profile. The utilized procedure req uired 3 different steps: 1) evaluation of chemical and sensorial profi le on a training set of 15 oils; 2) construction of a PLS 1 prediction model For each sensory variable on the basis of training set samples chemical profile; 3) prediction of sensory profile of a Further oil se t (test set) with a given chemical profile. The prediction model was t ested by comparing the predicted data with the experimental ones.