The Evolution of Community Structure in a Coauthorship Network

Mechanisms such as triadic closure and preferential attachment drive the evolution of social networks. Many models use these mechanisms to predict future links, and they generate realistic networks with scale-free degree distributions. These social networks also have community structure, or sets of vertices which are more connected to each other than the rest of the network. To study the evolution of research groups of scientists in a coauthorship network, we use a timeheterarchy representation to extend the mechanisms driving the evolution of the network to the level of this community structure. Specifically, we examine changes in the structure of groups in terms of mechanisms analogous to triadic closure and preferential attachment, and as a result, we find that the network evolves in the same way at the group-level and the individual-level. In addition, we find that interactions at the group-level might affect interactions at the individual-level in that members of a single group are more likely to strengthen their relationships than members of separate groups.