Using the adaptive fuzzy-neural system (ANFIS) in modeling storage conditions of grape fruits coated with maltodextrin containing potassium nanocarbonate and Pyracantha extract

In this research, due to the lack of a study on the modeling of the storage conditions of grape fruits coated with maltodextrin containing potassium nanocarbonate and pyracantha extract using the fuzzy-neural inference system (ANFIS), different storage times (zero to 60 days), concentration Potassium nanocarbonate (zero to 2%) and concentration of pyracantha extract (zero to 1.5%) used in edible coating as input and the amount of Brix, the percentage of dropping berry, the amount of malondialdehyde, the browning of the grapes cluster, aroma and taste of samples were considered as output. Three Gaussian, triangular and trapezoidal membership functions with 2-2-2 and 3-3-3 membership functions were investigated. The results showed that trapezoidal function with 3-3 membership function and Gaussian function with 3-3 membership function were selected as the best model for Brix output variable and percentage of dropping berry, respectively. finally, triangular model with 3-3 membership function was selected for cluster browning and malondialdehyde variables. On the other hand, the results indicated that with the increase in storage time, the amount of Brix, the percentage of dropping berry, malondialdehyde and cluster browning increased in all samples, but the amount of aroma and tast of the samples decreased. the use of coating containing potassium nanocarbonate and pyracantha extract caused the speed of these changes to decrease. In the end, it can be said that the high correlation coefficients between the laboratory results and the model outputs indicate the acceptable accuracy and usability of these models in controlling the storage conditions of grape fruits coated with maltodextrin containing potassium nanocarbonate and pyracantha extract.