Research On Optimized Fuzzy Classifier Based On WM And PSO

System modeling based on sample data is an efficient type of fuzzy classifier.The Wang-Mental(WM)method is an effective method to extract fuzzy rules from data directly without prior knowledge.The WM method is simple,efficient and practical.However,the algorithm is easy to extract the rules with lower reliability.Therefore,the fuzzy rule base which is extracted by WM method needs to be further optimized.The particle swarm optimization(ISPO)algorithm is a kind of iterative method based on the evolution.Application of ISPO algorithm in the field of fuzzy classifier is mainly to integrate the fuzzy knowledge base.Via optimizing by ISPO algorithm,the structure of the original fuzzy rules is changed into another with a better combination of efficiency.However,the coding process of this application is complex,and it is conservative in the fitness function.In this paper,intelligence single particle optimizer(ISPO)based on ISPO algorithm are used to optimize the rule base.Compared with the ISPO algorithm,the ISPO algorithm has faster convergence effect.The WM algorithm can extract the efficient implementation of rules,but the rules are lack of sample correlation,which leads to the classification accuracy of the rule base dropping.In order to avoid the influence of this phenomenon,the ISPO algorithm is used to optimize the rule base,and the degree of association is used to modify the rules in the fitness function.Moreover,the fitness function is positively correlated with the classification accuracy,which ensures the high accuracy of the rule base.Reverse thinking,ISPO algorithm has a major drawback which is that the initial population of ISPO algorithm is randomly generated,so as to reduce the speed of convergence of the algorithm,the initializing particle of the proposed method should be the rule base with a certain degree of accuracy.Based on the above analysis,this paper proposes an optimized fuzzy classifier WPFS algorithm based on WM and ISPO.Further analysis of the WM algorithm and the ISPO algorithm shows that the two algorithms have a high P-ability.Therefore,the algorithm is P-ized and reconstructed as P-WPFS algorithm.In order to verify the validity of the P-classifier model,the P-model and MapReduce model are combined to form the WPFS-MR fuzzy classifier model.The WPFS-MR fuzzy classifier model greatly improves the processing efficiency of the algorithm,so that the algorithm can give the classification results within an acceptable time range.At the same time,it can ensure the accuracy of the classification results to maintain high accuracy which solves the problem of large data classification.