Survival analysis for modeling critical events that communities may undergo in dynamic social networks

Abstract

In dynamic social networks, communities may undergo various changes over time. For example, a community may split into several other communities, expand into a larger community, or shrink to a smaller community, or several communities may merge into one community. This is an important and difficult issue in the study of social networks. In the current literature, there is a lack of formal approaches for modeling and predicting critical events over time. This prompted us to seek a new approach based on a more general method of event prediction in order to make use of past histories of change. To this end, we propose here a general risk model that utilizes survival analysis techniques to understand and predict future changes in the topological structure of dynamic communities. We illustrate the performance of the proposed method on real networks extracted from the DBLP and Yelp datasets.