Social Networks: Finding Highly Similar Users and Their Inherent Patterns

Social networks provide a framework for connecting users, often by allowing one to find his pre-existing friends. Some social networks even allow users to find others based on a particular interest or tag. However, users would ideally like the option of finding others who share many of their inter- ests, thus allowing one to find highly-similar, like-minded individuals whom we call clones. This functionality would not only be a strong improvement to current social networks, but it could yield interesting research work regarding graph theory and sociology. In this paper we explore a means for finding"clones"within a large, sparse social graph, and we present our findings that concern patterns of shared interests. Additionally, we explore how this correlates with connectivity and degrees of separation. With our introductory work, we hope to en- courage future work regarding developing ecient, accurate algorithms for finding clones. In addition, we aim to inspire others to further our data mining eorts toward understand- ing the relationship between connectivity and having shared interests.