RAIN: social role-aware information diffusion

摘要: Information diffusion, which studies how information is propagated in social networks, has attracted considerable research effort recently. However, most existing approaches do not distinguish roles that nodes may play the diffusion process. In this paper, we study interplay between users' and their influence on diffusion. We propose a Role-Aware INformation model (RAIN) integrates role recognition modeling into unified framework. develop Gibbs-sampling based algorithm to learn proposed using historical data. The can be applied different scenarios. For instance, at micro-level, used predict whether an individual user will repost specific message; while macro-level, use scale duration of evaluate real media data set. Our performs much better both micro- macro-level prediction than several alternative methods.