Research Of Clustering Analysis Based On Swarm Intelligence Optimization Algorithm

The swarm intelligence op timization algorithm is an optimization method to simulate the intelligent social behaviors of natural biological groups,and performs well in optimization.Compared with the traditional solution,the algorithm is easier to understand and realize,and it works effectively in solving large-scale and complex optimization problems,and has drawn much attention from domestic and foreign researchers.With the development of information technology and the extensive application of the database,it is very important for people to obtain the necessary useful information from the massive data,data mining has become a frontier research topic.Cluster analysis as a key branch module for data mining,which is an important research method for understanding and analyzing data,and widely used in engineering design,information management,market analysis and many other research areas at present,promote the continuous progress of human society.This paper,focused on the clustering analysis of swarm intelligence optimization algorithm.The particle swarm optimization algorithm and the cuckoo search algorithm were improved,and then applied to clustering analysis respectively combining K-means algorithm for optimizing initial clustering center selection.The experimental results show that the clustering effect is evidently superior to the traditional swarm intelligent algorithm and clustering algorithm,which can achieve better clustering application effect under certain conditions.This article is specified in the following aspects:At first,analyzed the current situation and basic theory of swarm intelligence optimization algorithm and clustering analysis,and introduced two classical algorithms including K-means clustering and K-medoids clustering.And then,it focused more on particle swarm optimization and cuckoo search algorithm,and applied all these to clustering analysis.Secondly,facing with the disadvantage that the particle swarm algorithm(PSO)was easy to fall into the local solution,this paper tried to add the adjacent disturbances factor in the particle swarm algorithm formula,and made the algorithm follow the adaptive optimization learning strategy,to enhance the global search capacity to avoid premature particles.In addition,as the initial clustering center had a great influence on the K-means algorithm,it optimized the initial clustering center based on the idea of ‘cluster degree'.Optimized initial particle selection and extended neighborhood search range,this paper proposed a simplified Particle Swarm Optimization for merging Adjacent Disturbances and K-means Initial clustering algorithm(ADPSO-IKM).The simulation experiments on classical datasets show that the ADPSO-IKM algorithm can accelerate the convergence speed,has higher accuracy with better stability.Thirdly,for sake of the problem that the late convergence rate of the cuckoo search algorithm(CS)was slow,the paper merged the idea of particle swarm optimization algorithm and followed to the adaptive optimized learning strategy.And introduced self-adaptive adjust step factor and dynamic change Pa,to equalize the global and local fine search capability of CS algorithm.And then,combined with the K-means algorithm based on ‘cluster degree' and distance equilibrium to optimize the initial center,this paper developed a Cuckoo Search for Self-adaptive Adjustment and Optimization of Initial K-Means clustering algorithm(CSSA-OIKM).The simulation results show that the CSSA-OIKM algorithm has a better clustering convergence effect.