Artificial Fish K-means Algorithm And Its Application In Communication Customer Segmentation Research

K-means algorithm for the processing of large data sets has good expansibility,high efficiency But K-means algorithm exist deficiencies,Its deficiency is the clustering effect is largely influenced by the initial clustering center selection,and overall cost easily affected by initial amount data.So,in order to make up for the traditional clustering algorithm in clustering analysis of some drawbacks.Some researchers use bionics,the mechanism of biological evolution from some intelligent fusion algorithm and clustering algorithm,Optimization clustering process,improve algorithm efficiency.While Artificial Fish Algorithm(Artificial Fish Swarm Algorithm,AFSA)as a kind of bionic optimization Algorithm,its has low requirements for initial parameters,simple structure and optimization speed,etc.its can be used in many of the solution of the optimization model.So the artificial fish algorithm is used to optimize K-means algorithm will be able to effectively improve the clustering effect.Based on thorough analysis of K-means algorithm and artificial fish algorithm based on the research status at home and abroad,in the further study of K-means algorithm and relevant theories and methods of the artificial fish algorithm.Aiming at the shortcomings of the K-means algorithm,With artificial fish algorithm to optimize K-means algorithm clustering process.By constructing artificial fish individual determination algorithm,the objective function,put forward to solve the problem of clustering artificial fish K-means algorithm,Paper design related experiments,through the programming will be artificial fish K-means algorithm,Through programming the artificial fish K-means algorithm,and K-means algorithm deal with the data set in the same conditions,using the optimal value and optimal number of values,and accuracy evaluation criterion,comparative assessment results and to verify this algorithm performance,At the same time the algorithm used in customer segmentation in the field of communications,formulate the corresponding marketing strategies according to the results of the experiment group.