Research On Dynamic Community Discovery Based On Incremental Analysis

In the real world,there are a lot of social networks, such as friend networks, networks of scientists, information networks, and so on. In recent years, there was such as such as Facebook, Renren and Blog newly formed social networks. In these networks, there is often a relationship between some of the nodes more closely the relationship and some of the nodes while relatively sparse phenomenon, the structure of these nodes forming a close relationship is called community structure. Master network of community structures is of great significance for us. Community structure of our social networks to develop more problems with social awareness strategy can produce a lot of useful information; however, understanding community structure in social networks is very difficult, especially in dynamic social networks.Most of complex networks often change over time, in the constantly changing, community found of dynamic network to have a great challenge in the current dynamic network partitioning algorithm of community, in general, there were two ways. One is a static network all the time slice sequence combined into only a static network, and then only in a static network was subjected to community discovery; the other is static on the network each time slice communities were found. In fact, the network is not static, dynamic nature is the nature of the network, in this dynamic process, there will be some community structure not changing much, if every time you want to know the network of community structure, we work out individuals of the entire network so that have a great time complexity, in such a large social networks,which is content with our requirements for rapid detection information.In this paper, because of the above shortcomings, the author fully research community mining algorithms, especially found a lot of research work on the basis of direction of dynamic social network community,presents a dynamic community-based incremental discovery algorithm analysis CFIA (Community Finding on Incremental Analysis). It is used to identify the dynamic online social network community structure. In this way, a series of changes in the network after it in accordance with the previous network and the snapshot and the incremental change in consideration of the impact of the lower adjacent node by updating the situation quickly and effectively in the network community structure. In order to achieve a higher timeliness, we introduced the concept of incremental network, the network topology incremental analysis to determine the current structure of the network of community division. By actual dynamic network IkeNet-em6ail-long.net dataset experiments, and in the community and on the quality of the time division traditional MKBCD, IC algorithm are compared. MKBCD algorithm modules have a higher degree of accuracy Q and AC, but the time is too long, and IC algorithm is shorter, but the modularity Q and accuracy AC relatively low, CFIA algorithm results in divided communities modularity Q and accurate AC aspects degrees higher than IC algorithm close MKBCD algorithm, at time T on higher IC algorithm but far below MKBCD. The comparison of the results can be shown:In the case of the quality assurance division of the community, at the same time,the community reduces the time required to divide the use of community-based dynamic analysis of incremental discovery algorithm CFIA,which is found the accuracy and timeliness of community structure in dynamic networks.The innovation of this paper is that considering the incremental impact on neighbors community belonging, ensuring the quality of the community divided by incremental changes in the history of the network and dynamic network to discover communities, avoiding the high complexity of the the entire dynamic of all network nodes divided in every moment, greatly reducing the analysis based on the incremental time complexity and improve the timeliness of the community divided.