Research And Application Of Fuzzy Neural Network Based On Genetic Rule Clustering

With the development of science and technology,with the maturation and wide application of artificial intelligence technology,the fuzzy neural network model and the genetic algorithm optimized attract much attention where the fuzzy neural network mode integrating the advantages of neural network and fuzzy control and the genetic algorithm optimized by simulating the genetic laws of the biological world are also affected.In the classical fuzzy neural network model,the main problem is that when there are many variables in the sample and there are too many fuzzy subsets in each variable,it is very easy to cause the explosion of the rule combination,so that the model rules are too many to show one by one.This greatly increases the scale complexity and modeling time of the model.In response to this problem,many application processes rely on subtractive clustering based on real variables to extract representative classification rules to avoid the large number of rules to implement the classification process.This method actually turns the fuzzy problem into reality.Therefore,for the classification of fuzzy neural networks,whether there is a representative classification rule extraction based on fuzzy variables,in order to reduce the number of most fuzzy combination rules,to achieve an obvious fuzzy classification under the nonlinear mapping problem?No relevant research literature has been found yet.In this paper,aiming at the above problems,a nonlinear fuzzy neural network model based on fuzzy variable input and genetic selection rules is proposed.The relative membership degree of fuzzy variables is expressed as the chromosome(individual).Through the combination of K-means clustering and genetic algorithm,the clustering number K and the clustering center are genetically optimized,and the fuzzy class and fuzzy classification rules represented by the clustering center,thereby obtaining the Koptimum rules;Based on this,for any Input the sample,apply K fuzzy rules to classify,and combine the fuzzy network for nonlinear approximation.When the inputs and outputs are all fuzzy variables,a four-layer fuzzy neural network model is established using a small number of representative rules and nonlinear sigmoid function functions.With the aid of SPSS and MATLAB software,the fuzzy neural network method proposed in this paper is applied to study and analyze the fuzzy data of the comprehensive evaluation of university asset management performance.Satisfactory results are obtained and the effectiveness of the method is demonstrated.