Research Of Module Portion Methods Based On Particle Swarm Optimization Algorithm And Fuzzy Clusterin

With the rapid development of information era,information classification has become an indispensable method to get useful information quickly and efficiently in all walks of life.As one of the important ways to deal with classification problem,the fuzzy clustering analysis,which is applied in the module partition,has important theoretical significance and practical value.Because of sensitive to the initial value,Fuzzy clustering algorithm is often combined with optimization algorithm.But optimization algorithm has the drawback of the tendency to be entrapped in the local minimum,so this paper mainly do the following three aspects.At first,we analyze several algorithms,which are commonly applied in module partition,including the fuzzy c-means clustering algorithm,simulated annealing algorithm,fuzzy c-means algorithm based on genetic algorithm.Summarizing these algorithms' advantages and disadvantages,we propose to improve the particle swarm optimization algorithm,and get a based on improved particle swarm algorithm of the kernel fuzzy c-means clustering algorithm.Second,we aim at the limitation of particle swarm optimization,study and analyze the related literature,apply the mechanism of adaptive inertia weight,the weighted center as well as the thoughts of Metropolis accept standards to the algorithm in order to improve the particle swarm optimization algorithm.And then we have done the experiments to prove that the improved algorithm has more advantages than before.Thirdly,this paper proposes a based on improved particle swarm algorithm of the kernel fuzzy c-means clustering module partition method.The combination of FCM and parameter optimization algorithm is the most commonly applied in the module partition.Compared with ant colony algorithm and genetic algorithm,particle swarm optimization algorithm has the less parameters to adjust,and the convergence speed and performance are better than the former.Because the data deal with in module partition has a not high complexity,the advantage of quantum particle swarm optimization,which is the efficient processing of complex data,does not put to good use.So the combination of FCM and particle swarm optimization algorithm is more widely applied.We use the improved particle swarm algorithm to optimize the clustering number,which is combined with KFCM.And the results of simulation experiment have proved that the based on improved particle swarm algorithm of KFCM clustering method of module partition has much more superiority.And we also apply this method into the actual data processing module,which verifies its certain application value.