Modeling microblogging communication based on human dynamics.

Jianjun Xie,Chuang Zhang,Ming Wu

DOI: https://doi.org/10.1109/FSKD.2011.6020045

2011-01-01

Abstract:Microblog is a large-scale information sharing platform where intensive communications are taking place through interactive user behaviors. Previous studies have analyzed and modeled a series of traditional communications, such as letter, email and phone calls. The timing of human activities tends to be non-Poisson with bursts and heavy tails. Hence, several models have been proposed to explain human dynamics of bursts and heavy tails in various fields. However, as a newly developed product of Web2.0, microblog possesses inherent new characteristics, e.g. user-centric broadcast medium and asymmetric user relations. The communication in microblog is still poorly understood. Our work proposed an interest-driven model to simulate basic user communicating behaviors and processes. We came to the conclusion that, in microblogging communication, individual behaviors are bursts of rapidly occurring events separated by long periods of inactivity. Collective behaviors follow heavy-tailed Power Law distribution, whose origin is the descent of interests. Empirical statistics were also given to verify the model. © 2011 IEEE.