96 / 1971-01-01 00:00:00

Competition-based Information Propagation In Online Social Networks

5170,5171,5172,5173

In this paper, we propose a competition-based two-stage information propagation model for multiple topics. Unlike the previous work that focus on analyzing network structure and user profiles, we study the information propagation from a perspective of topic competition and treat it as two stages: gaining and losing
attention of the participating users. In order to better understand the dynamics of topics, we investigate the data from Sina Weibo. The actual observations not only provide evidence for our model assumptions, and help us to set experimental parameters to validate our model. Simulation results demonstrate that this model could effectively generate the single-peak and heavy-tail phenomenon in the propagation of information and is consistent with the actual observations. By means of simulations, we explore the effects of topic arrival time and user defection on the dynamics of topics in our model. We find that a single topic is unlikely to be a burst one when there are more competitive topics.