Propagation Modeling and Analysis of Incidental Topics in Blogosphere

Abstract

Blog has become one of the most important media among the general public, and the propagation modeling of incidental topics in blogosphere is of great interest in social network studies. Most existing analysis methods are based on the infection models in epidemiology. However, many of these models are inconsistent with the widely observed power-law decay of the propagation velocity. In this paper, the propagation of incidental topics is described by a susceptible infection (SI) model based on the individual fitness. It is proved that the propagation velocity will asymptotically drop with power-law if the fitness density function satisfies certain conditions. Moreover, if the individual fitness is of uniform distribution, analytical solution of propagation velocity can be obtained based on our model. Model verifications are performed on the data from several widely discussed popular topics in Sina Blog and the results show that our model is consistent with the actual propagations.