Ulrik Brandes / Michael Hamann

On the Persistence of Strongly Embedded Ties

[PDF herunterladen]

The identification of cohesive groups that are stable over time is one of the main concerns in dynamic clustering. Typically, this is achieved by incorporating previous information. In this chapter, we show that strongly embedded ties, which make up the backbone of a network, not just allow to distinguish between intra- and inter-cluster ties but, more importantly, are indeed relatively more persistent over-time. This points to a new definition of cohesive groups in dynamic settings that do not only take previous but also future information into account.