Cultivar discriminating ability of ANNs on Apulian monocultivar EVOOs was studied.
Merceological, NIR and 1H NMR data were used as ANNs training sets.
ANN models based on NMR data showed the highest accuracy in classifying cultivars.
The most information about cultivars was contained in very few NMR peaks.
Performance was not influence by the milling method nor the crop year.