A Unified Information Diffusion Model for Social Networks

Abstract

Modeling information diffusion for social networks enables analyzing and predicting the impact on a specific topic or event. Most existing information diffusion models only take information influence or diffusion among the network users independently into consideration, which can hardly depict the diffusion in a booming large-scale and complex social network. In this paper, we propose a unified information diffusion model which jointly combines information influence and the diffusion process. Based on the unified model, most existing diffusion models can be reconstructed and implemented easily by setting influence and diffusion functions. In this paper, we use this model reconstructed five traditional models and we evaluate the model with a real collected dataset of 2.2 million blogs and the results show that the unified model can be adopted easily to fulfill classic diffusion models. Furthermore, the model sheds light on designing novel explainable information model in the future.