Study of Neural Networks and Its Application based on Fuzzy Adjustment

With the development of fuzzy and neural network theory, the role both play in engineering applications is also growing, but both their disadvantages are also gradually revealed. The neural network has a strong ability to learn, but it is a typical "black box" model, the knowledge acquired in connection reserves the right, the user can not directly understand and use; fuzzy model of human natural language description of the problem, using fuzzy rule sets are derived, and therefore easy to understand, but it can not be expert knowledge or a large number of sample data directly into the inference rule base, and its lack of ability to learn and hard to improve itself. How to combine the advantages of both organically each other, it is increasingly becoming an issue of concern. Aiming at this problem, we made a study in the theory and application of neuro-fuzzy fusion.