Cognitive learning in neural networks using fuzzy systems

Adaptation and recognition problem in multipurpose environments is always called as one of the most important and useful problems between the recognition algorithms and mechanisms. One popular approach is using mechanisms which are trying to train the recognition system on a targeted pattern, to concentrating all the capacity of recognition system onto recognition or prediction process. Cognitive structures are among these intelligence learning solutions. Since cognitive structures used in living organisms have shown their eligibility on recognition and classification with appropriate accuracy, utilizing of these structures is rational. This paper proposes three cognitive principles based on neural network as a universal approximator. First a central control unit, a fuzzy system, will be shown to determine the adaptation rate for a target sample. Second, a minimal optimization will be performing for each sample. Third, a recognition system is going to use for recognition in sort of different environments.