Research On Community Discovery And Evolution In Complex Networks

With the rapid development of Internet, the Web, which have been a lot of communities. The communities can represent the Web's social activities. Web will be organized into communities which allow users to understand the knowledge of information and the development trend of organizational structure of the Web, and can achieve the purposes of organizing the network level by identifying and distinguishing the communities. The technology of community identification that becomes a research hotspot.In practice, existing community identification methods are being two major problems to be solved. First, considering only the community connections between nodes, while ignoring contents of the entities represent by nodes; second, ignoring the Dynamic of the network. To overcome the above shortcomings, proposing a new algorithm for community identification, and analyzing the community evolution. The main work includes the following aspects:First of all, taking into account the different pages in the Internet express different contents, the paper proposes an algorithm for community identification based on the Web pages contents similarity and the link relation between the Web pages. It uses the improved cosine similarity formula to calculate the similarity of the pages and the themes. Considering the links between pages, we can get the communities both in hyperlinks and contents, which keeps the integrity of the communities.Secondly, there may be some nodes in the community, which the links between them more closely than the other nodes, thus forming a higher compactness of small communities, the community appears hierarchy. To this feature, this paper introduces the concept of community hierarchy to make the community into several layers based on the compactness of nodes, and analyze the hierarchy of the community. Finally, to define the types of the evolution that may appear during the evolution of the communities, and give the metrics of the evolution. Analyzing the evolution of communities based on the hierarchy of the community. It can reflect the complexity of the themes'changes in the evolution of communities, analyzing the community evolution in the Complex Networks.Experimental results show that the algorithm can find the community structure in the Complex Networks efficiently, and reflect the evolution of the community.