Modeling by artificial neuron networks (ann) of drying (under hot air) of cashew apple (anacardium occidental l.)

The cashew apple is a product of proven nutritional importance and perishable due to its high water content. A reduction of this would allow its conservation and can be done by osmosis-drying in order to limit the loss of nutrients. This drying study aimed to describe the behavior of water outflow in cashew apples treated with molasses and sucrose solutions in order to predict their final water content. Drying experiments were carried out on fresh cashew slices in an oven at 50°C by static gravimetric weighing. The drying data was used to create and simulate the evolution of the water content from an artificial neural network (multilayer perceptron). The results indicate that the 4-2-1 structure artificial neural network exhibited the best abilities to predict and simulate the end-of-drying water content of cashew apples treated with R² > 0.9999 and EQM (9.7252 E -09) <10-9.