Modelling topic propagation over the Internet

Because of the booming of the Internet, content security is becoming more intractable, because of the emergence of complex contents and the diversity in human activity on the Internet. The article proposes a model for the dynamics of topic propagation over the Internet. Topics on the Internet are considered as clusters of contents on websites, which describe various kinds of events. The model accounts for the behaviours of websites, such as anti-infection ability, recovery ability, spreading ability and effective propagation rate. A new topic diffusion mechanism incorporating Markov model based on topic activity transition is employed in the model. By means of simulations, we explore the time-dependent spreading of topics in directed scale-free networks, in which nodes are considered as websites and directed links represent the source dependencies between websites. The simulation results accord with the actual observation very well.