Estimation of papaw (Carica papaw L.) moisture content using adaptive neuro-fuzzy inference system (ANFIS) and genetic algorithm-artificial neural network (GA-ANN)

Adaptive neuro-fuzzy inference system (ANFIS) and genetic algorithm-artificial neural network (GA-ANN) were used for modeling of the hot-air drying kinetics of papaw slices. The ANFIS and GA-ANN were fed with 3 inputs of drying time (0-320 min), drying temperature (40, 50 and 60 °C) and slice thickness (3, 5 and 7 mm) for prediction of moisture ratio (MR). The triangular membership functions (MFs) were applied and 27 rules were provided for the ANFIS designing. The developed ANFIS predictions were close to the experimental data (R2 = 0.9967 and RMSE = 0.0161). The optimized GA-ANN, which included 7 hidden neurons, predicted the MR with a good precision (R2 = 0.9936 and RMSE = 0.0220). The effective diffusivity for papaw slices was within the range of 6.93 ×10-10 to 1.50×10-9 m2/s over the temperature range. The activation energy was found to be 32.5 kJ/mol indicating the effect of temperature on diffusivity.