Prediction of Density of Fruit Juice Using Neural Networks as Function of Concentration and Temperature

Abstract

An artificial neural network (ANN) model was used for the prediction of density of fruit juice as a function of concentration and temperature. The various fruit juices considered were peach juice, orange juice, pear juice and malus floribunda juice. The density data used during modeling were taken from the literature for a wide range of concentration (10-71degree Brix) and temperature (0-80degree C). ANN topologies were evaluated while developing the optimal ANN model. The optimal ANN model consisted of two hidden layers with four neurons in the first and three neurons in the second hidden layer. This model was able to predict density with a mean sum square error of 0.0004 g²/cm6.