Comparison of kinetics, neural network and fuzzy logic in modelling texture changes of dry peas in long time cooking

Kinetic, neural network and fuzzy logic models were proposed to model the t extural changes of dry peas cooked at 70, 80, 90 and 100 degrees C for up t o 240 min. The results were compared to the first order kinetic and Rizvi a nd Tong (R-T) models. It was observed that the textural changes in cooked p eas vs. time did not follow the first order reaction kinetic model. The neu ral network model consistently produced the best fit to the experimental da ta, and the first-order-reaction kinetic model the worst. The performance o f the other three models, that is, the proposed kinetic, fuzzy and R-T, var ied. The models were also validated and a similar pattern was observed. Com pared to the traditional kinetic models, the neural network and fuzzy logic models are more flexible. They use the weight matrix (NN model) oi members hip function and fuzzy rules, instead of activation energy and constant.