Distributed Adaptive Neuro Intuitionistic Fuzzy Architecture for prediction of the dose in gamma irradiated milk products

In this paper, a Distributed Adaptive Neuro Intuitionistic Fuzzy Architecture (DANIFA) with a second order Takagi-Sugeno inference is presented. The architecture represents a layered set of simple fuzzy inferences connected in a distributed way, thus minimizing the number of the interconnected fuzzy mles and their associated parameters. The flexibility of the designed structure to handle uncertain data variations is complemented, by embedding an Intuitionistic fuzzification approach. A simple two-step gradient descent algorithm with a fixed learning rate is used as a learning algorithm of the proposed architecture. To test the prediction abilities of the designed model a biological case for estimation of the low gamma irradiation dose to destnict the protein fractions in milk products with potential uncertain data variations is studied.