Identifying Structural Hole Spanners in Online Social Networks Using Machine Learning

Online social networks play an important role in our daily activities. As an important concept in social network analytics, the structural hole theory shows that the positions in social networks that can bridge different user groups will get benefits. Existing solutions for identifying structural hole spanners normally require the knowledge of the entire social graph. In this paper, we propose a novel solution to uncover structural hole spanners according to the users' profiles and user-generated contents (UGCs), instead of referring to the entire social graph. We propose a machine learning-based model to implement the identification. We further leverage the ego networks and the cross-site linking function to enhance the identification. A real-world dataset collected from Foursquare and Twitter is used to evaluate the identification performance of our model. The results show that our model can achieve a high F1-score of 0.857.