Modelling the Influence of Origin, Packing and Storage on Water Activity, Colour and Texture of Almonds, Hazelnuts and Walnuts Using Artificial Neural Networks

The present work assessed the influence of different factors on some physical and chemical properties of nuts. The factors evaluated were the presence or absence of the inner skin, geographical origin, storage conditions (ambient temperature, in a stove at 30 and 50 °C, in a chamber at 30 and 50 °C and 90 % RH, refrigerated and freezing) and type of package (none, low density polyethylene and low density polyethylene). The fruits studied were almonds, hazelnuts and walnuts from different countries. The properties measured were moisture content, water activity, colour coordinates (L*, a* and b*) and texture parameters (hardness and friability). Experimental data were modelled using neural networks. The results showed that the almonds from Spain and Romania had aw greater than 0.6, and therefore, its stability was not guaranteed, contrarily to the other samples that presented values of aw lower than 0.6. The colour coordinate lightness varied from 40.60 to 49.30 in the fresh samples but decreased during storage, indicating darkening. In general, an increase in hardness and friability was observed with the different storage conditions. Neuron weight analysis has shown that the origin was a good predictor for moisture content and texture; whereas, the storage condition was a good predictor for aw and colour. In conclusion, it was possible to verify that the properties of nuts are very different depending on origin; they are better preserved at lower temperatures and the type of package used did not impact the properties studied.