The Research On Clustering Algorithm Based On Granular Compution And Rough Set Theory

The rapid development of global information technology and Internet,making people have more and more demands for sharing network resource.And lots of data information caused the data inflation and information explosion.There is an urgent need for people to find a scientific and reasonable way to get effective and reliable information from a mass of mixed data.Data mining is an effective approach to solve this problem.It can help people make the right and efficient decision after professional processing of specific data.Cluster analysis is an key content of data mining,becoming the research topic of many experts and scholars.Based on the classical clustering method,this paper analyzes the limitations of clustering algorithm,and then studied the swarm algorithm,particle swarm optimization algorithm,rough set and granular computing theory.Finally,combining the colony,particle swarm,rough set and granular computing to optimize the traditional clustering algorithm.The main work is as follows:(1)Due to the disadvantages of traditional K-medoids clustering algorithm such as randomly select the initial clustering center,low accuracy and the poor global search ability,this paper proposed a K-medoids clustering optimization algorithm based on artificial bee colony.The proposed algorithm is combined with improved grain computing and maximum distances product method to select the initial clustering center,dynamically adjusted the search step length with iteration increasing,and then introduced the selection probability based on sorting instead of depending on fitness directly to choose follow bees,further speed up the convergence speed of the algorithm,avoided the premature convergence.The experimental results show that the algorithm can reduce the dependence on the initial clustering center,has high accuracy and strong stability.(2)To deal with the deficiencies of the traditional K-means algorithm of its depending overly on the selection of the initial clustering centers,low efficiency of handling boundary objects,low clustering accuracy and poor stability,this paper proposed a new algorithm based on particle swarm optimization and rough set.Firstof all,the density and maximum distances product method is used in this approach to achieve initialization.Then the method of combining linearly decreasing and random distribution is adopted to produce inertia weight.By adjusting the learning factor dynamically and introducing random particles,to increase the diversity of population.At last,the rough set and improved particle swarm optimization are combined to optimize K-means.The results of the simulation of several commonly used UCI benchmark datasets show that the new approach can not only reduce the dependence on the initial cluster center and deal with the boundary data effectively,but also obtain high accuracy and strong stability.