Research On Community Partition Algorithms Based On Cloud Aggregation

Complex network is common in nature, including social network, technology network, biological network and so on. Finding the community structure in complex network has a profound significance. However, the existing community division algorithms have disvantages in analyzing node attributes, link weight, evaluation criterias and applicability aspects, which can not be used for all social network, so a special algorithm is proposed in this paper.First of all, it need to meet technical index in the social network subject to area, that is, regional continuity, node attributes similarity, neighbor nodes sharing, dynamic adaptability and community scale equilibrium. Secondly, the period of water vapor forming clouds resemble community nodes aggregating, so community division algorithm based on cloud aggregation is put forward, which can be used in both static and dynamic community discovery algorithms. Finally, the algorithm is applied to static urban community and dynamic self-driving social network community by the platform of Matlab. This algorithm got higher modularity than that of the similar community division algorithms, which proved the validity and superiority of the algorithm. The experimental results show that the proposed algorithm in this paper can improve modularity in the community structure.