Efficient community maintenance for dynamic social networks

Community detection plays an important role in a wide range of research topics for social networks. The highly dynamic nature of social platforms, and accordingly the constant updates to the underlying network, all present a serious challenge for efficient maintenance of the identified communities-How to avoid computing from scratch the whole community detection result in face of every update, which constitutes small changes more often than not. To solve this problem, we propose a novel and efficient algorithm to maintain the communities in dynamic social networks by identifying and updating only those vertices whose community memberships are affected. The complexity of our algorithm is independent of the graph size. Experiments across varied datasets demonstrate the superiority of our proposed algorithm in terms of time efficiency and accuracy.