||A Study on Finding Core Communities in Social Networks
In our society people are integrated into communities to share ideas, information and to achieve a common goal. Within a community, a small group of individuals holds great popularities and leaderships, which we call a core community. A core community takes general tasks, coordinates the interactions among members, and holds the important source of the social norm, values and subcultures. Therefore, identifying core communities is of great importance to understand our social systems. There has been concert effort to find core communities embedded in social networks. However, one limitation is that they are too coarse to find multiple cores in a network. We, therefore, proposed a novel communitymodel called dense core. It can capture almost perfect subgraphs and also well-separated subgraphs. On the basis of this unique nature, we have shown its effectiveness in identifying multiple core/periphery structures. The spontaneity is supported by the fact that the dense core model is relevant to various community models in different traditions such as pseudocliques and graph partitioning. We applied the dense core model for visualizing a graph to find multiple core communities. In addition, we applied our model for the change analysis of dynamic social networks and illustrated its effectiveness.
Hokkaido University（北海道大学）. 博士(情報科学)