A Model of Information Diffusion in Interconnected Online Social Networks

作者：Gaeta Rossano^*

来源：ACM Transactions on the Web, 2018, 12(2): 13.

DOI：10.1145/3160000

• Online social networks
• information propagation
• information diffusion
• information spreading
• data diffusion
• data propagation
• data spreading
• interconnected networks
• multiple online social networks
• generalized random graphs
• online social internetworking
• mathematical modeling

Online social networks (OSN) have today reached a remarkable capillary diffusion. There are numerous examples of very large platforms people use to communicate and maintain relationships. People also subscribe to several OSNs, e.g., people create accounts on Facebook, Twitter, and so on. This phenomenon leads to online social internetworking (OSI) scenarios where users who subscribe to multiple OSNs are termed as bridges. Unfortunately, several important features make the study of information propagation in an OSI scenario a difficult task, e.g., correlations in both the structural characteristics of OSNs and the bridge interconnections among them, heterogeneity and size of OSNs, activity factors, cross-posting propensity, and so on. In this article, we propose a directed random graph-based model that is amenable to efficient numerical solution to analyze the phenomenon of information propagation in an OSI scenario; in the model development, we take into account heterogeneity and correlations introduced by both topological (correlations among nodes degrees and among bridge distributions) and user-related factors (activity index, cross-posting propensity). We first validate the model predictions against simulations on snapshots of interconnected OSNs in a reference scenario. Subsequently, we exploit the model to show the impact on the information propagation of several characteristics of the reference scenario, i.e., size and complexity of the OSI scenario, degree distribution and overall number of bridges, growth and decline of OSNs in time, and time-varying cross-posting users propensity.

• 出版日期2018-6