Application Of Particle Swarm Optimization Algorithm In Data Clustering

Cluster analysis is a very important research direction in the field of data processing.The current clustering algorithms have their own advantages,but how to cluster the data with complex distribution is still an important issue in the current research.In this thesis,a hybrid particle swarm clustering algorithm is proposed by combining particle swarm clustering with spectral clustering.Hybrid particle swarm spectral clustering algorithm retains the characteristics of hybrid particle swarm and spectral clustering algorithm.The specific work is as follows:Firstly,the clustering problem is regarded as the optimization problem of finding the minimum fitness,the random initialization ability of particle swarm algorithm is used to escape from the local optimum,and the spectral clustering principle is used to process the data in advance,and a new hybrid particle swarm spectrum clustering algorithm is proposed to explore the problem of improving the clustering accuracy.Second,the hybrid particle swarm spectrum clustering algorithm is deeply studied from the three aspects of determining the number of clusters,reducing the complexity of the algorithm and screening the interference points.Through the principle analysis and numerical test,the better operation strategy of the three aspects is determined.Thirdly,through five quantitative indexes of clustering evaluation,numerical experiments are carried out on the proposed algorithm,which is compared with other four algorithms,and the advantages and disadvantages of the proposed algorithm are analyzed.