کارآیی زئولیت و اسید سیتریک در کنترل رشد کپک و تولید آفلاتوکسین در ضایعات نان های خشک در سطح شهر مشهد و مدل سازی آن به روش شبکه های عصبی مصنوعی

In this research, the modeling with Artificial neural network and Multilayer – Perceptron were used in order to evaluate the zeolite and citric acid's usage in reducing of Aflatoxin's production in stale dry breads across the Mashhad. Since, the stale breads are the main sources of the livestocks's feeding, and because of the availability of proper environmental conditions for growing molds, these breads are severely contaminated by mycotoxin and especially Aflatoxin, and this make some anxieties about the human being and animal's life. So, the foodstuff's contamination by mycotoxin, should be controlled accurately through the food chain. The results have shown that modelling with ANN is a suitable method especially in food industries, and also the addition of zeolite as compared with Citric Acid, cause the َAflatoxin to reduce more. In this manner, the interaction of zeolite and citric acid caused the Aflatoxin to decrease more, than when zeolite or citric acid are being used alone. Based on these results, Artificial neural network model for zeolite with one hidden layer, hyperbolic tangent function as the transfer function, Levenberg-Marquardt method as the learning rule, 3 hidden neurons, %60 for training subset and %20 for each of validation and test subsets with the correlation coefficient 0/973 had the best overfiting. The modeling results indicate that there is an excellent compatibility between the experimental and predicted values of Aflatoxins.