Reconstruction of social group networks from friendship networks using a tag-based model

• A tag-based model is proposed to explain the mechanism on the growth of social groups.
• In the model, social groups expand on a friendship network based on users’ tags of interest.
• Users’ activity in joining group is related to their degree of friendship network.
• Various distributions of the simulated group network are in agreement with empirical findings.
• Our model throws light on the reconstruction of institute-based relationships.

Social group is a type of mesoscopic structure that connects human individuals in microscopic level and the global structure of society. In this paper, we propose a tag-based model considering that social groups expand along the edge that connects two neighbors with a similar tag of interest. The model runs on a real-world friendship network, and its simulation results show that various properties of simulated group network can well fit the empirical analysis on real-world social groups, indicating that the model catches the major mechanism driving the evolution of social groups and successfully reconstructs the social group network from a friendship network and throws light on digging of relationships between social functional organizations.