Dynamics of group cohesion in homophilic networks

Understanding cohesion and homophily in empirical networks allows us build better personalization and recommendation systems. This paper proposes a network model that explains the emergence of cohesion and homophily as an aggregate outcome at the group- and network-level. We introduce two simple mechanisms that capture the underlying tendencies of nodes to connect with similar and different others. Our main theoretical result presents conditions on the network under which it reaches high degrees of cohesion and homophily.