Estimation of Equilibrium Moisture Content of Pistachio Powder through the ANN and GA Approaches
Abstract In this study, two intelligent tools of genetic algorithm (GA) and artificial neural network (ANN) were employed to use experimental data to predict equilibrium moisture content (EMC) of Persian pistachio powder. Initially the moisture sorption isotherms of pistachio powder were determined by gravimetric method at different temperatures (15, 25, 35 and 40°C) and constant relative humidity’s (0.11, 0.23, 0.36, 0.49, 0.62, 0.75 and 0.88 aw values) and then traditional mathematical models including BET, Iglesias and Chirife, GAB, Caurie and Freundlich were used to check the fitness of experimental data. Later the experimental data were compared with similar data obtained from GA and ANN models. The overall results showed that the Caurie model had high performance to predict EMC and revealed that GA model had greater accuracy to predict EMC of pistachio powder with very high R2 values (equal to 0.9996).