Dynamic clustering for social networks based on evolutionary computation

Abstract

Social networks are all around us and these networks are dynamic and time-evolving in nature. However, most current research focuses solely on the a possibly complete social network or a subset of that network. Thus, there is a lack of research in the field of dynamic network clustering.

Dynamic network clustering requires more attention since social networks are dynamic and any minor changes to its network alters the structure entirely. Currently, most research focuses on the clustering of static social networks where the network is either a small subset of a huge network structure or the entire network structure as a whole. Thus, since social networks change over time, it is more impactful to investigate the clustering problem in the context of dynamic social networks.

The intention of this project is to incorporate evolutionary computation, which is a nature-inspired algorithm in the context of dynamic social network clustering. Several social networks, with recorded time stamps, will be used to depict a growing social network. The optimal clusters are determined with each given time stamp. Each time stamp will provide a fairly different cluster structure and a possibly different cluster count.