Multi-message topic dissemination probabilistic model with memory attenuation based on Social-Messages Network

Current researches give priority to the diffusion of single message, but the diffusion of multi-messages at the same time in the actual network also exists. The diverse correlation of the messages will influence each other in the diffusion. It should be taken into consideration. This paper works to make a definition to the framework of Social-Messages Network. Based on it, a multi-message topic dissemination probabilistic model with memory attenuation is put forward, which introduces the correlation among messages. We adopt a simple learning strategy to gain the diverse correlation of messages. Then, the numerical simulation is utilized to analyze the model, whilst the relationship of the model parameter with the scope of the topic diffusion and the spread speed are studied and analyzed. With the related discussion data on Twitter, an empirical study is made to the model and the diffusion progress of the message is anticipated, which suggested that the anticipation is fundamentally in line with the actual data, and the estimated value of our model is closer to the reality than the classic diffusion model. Study on the topic diffusion will be conducive to the understanding and the anticipation of the multi-messages spread.