Modeling and Estimation of Banana Slice Browning during Drying using Artificial Neural Network.

In this study two primary models were used for mathematical modeling and artificial neural networks were used for estimation of the browning behavior of banana. Comparing between the models was shown that the first-order model (R-square close to 0.98) had presented the browning behavior of banana slices better than the zero-order model. The feed-forward ANN with Levenbery-Marqwardt algorithms as learning role, tangent sigmoidal as transfer function, 20 neurons in first hidden layer and 10 neurons in second hidden layer had the best estimation due to R-square of 0.9813 and RMSE of 0.6403.