Department of Computer Science
Stochastic network motif detection in social media
Network motifs refer to recurrent patterns of interconnections which are found to be over-represented in real networks when compared with random ones. Such basic building blocks can well characterize the structure of complex networks. Extending the building blocks to stochastic ones allows for more robust motif detection networks which are stochastic in nature. Network motif analysis, commonly adopted in bioinformatics, has recently been applied to also online social media. In this paper, we propose to detect stochastic network motifs in social media with the conjecture that social interactions are of stochastic nature. In particular, we apply a stochastic motif detection algorithm based on the finite mixture model to both synthesized datasets and real online datasets to evaluate the effectiveness. Also, we discuss how the obtained stochastic motifs could be interpreted and compared qualitatively with some of the results obtained from others which are recently reported in the literature. © 2011 IEEE.
Expectation-maximization algorithm, Mixture model, Social networks, Stochastic network motifs
Source Publication Title
Proceedings: 11th IEEE International Conference on Data Mining Workshops
Liu, Kai, William K. Cheung, and Jiming Liu. "Stochastic network motif detection in social media." Proceedings: 11th IEEE International Conference on Data Mining Workshops (2011): 949-956.