Research On Fuzzy Data Mining Based On Particle Swarm Optimization Algorithm

Clustering analysis is an effective data mining technology to identify patterns in data-sets.With the upgrading of science and technology and the continuous improvement of human needs,fuzzy clustering analysis can effectively deal with the large data-sets and more accurately describe uncertain relationship between patterns,thus become a hot research field of data mining.Among them,fuzzy c-means(FCM)clustering has become one of the most widely used fuzzy clustering algorithms because of its easy implementation,low memory consumption,fast speed of running and effective processing of large data sets.It has been applied in many fields,such as life science,business management,eco nomic finance,geological astronomy,especially in image segmentation and recognition,plays an irreplaceable role.However,the FCM algorithm also has the shortcomings of initial value sensitivity and easy to fall into local optimum.In order to solve the above problems,this paper hopes to provide effective help through joint intelligent algorithm on the basis of previous studies.This paper deeply studies the particle swarm optimization algorithm(PSO)in the intelligent optimization algorithm and FCM algorithm in the fuzzy clustering algorithm,and fuses them together,thus proposes the fuzzy clustering algorithm based on improved partic le swarm optimization algorithm(ESPSO-FCM).The main contents of this paper include: firstly,the particle swarm optimization algorithm is improved,and the random search step levy flights and PSO algorithm are repeatedly invoked by introducing the iterative model of eagle strategy.After the improvement,ESPSO the algorithm can get a balance between the global search and local development,and the particle has more satisfactory search ability and convergence speed.Then the improved PSO algorithm is integrated into FCM algorithm,solve the problem of FCM clustering algorithm and improve the clustering effect.Finally,the a lgorithm is applied to the actual data,and compared with other algorithms through simulation experiment,the experimental results show that the proposed algorithm is better than the traditional clustering algorithm.