Research On Agent-based Modeling Of Information Diffusion In Social Network

Research on information diffusion in social network is of both theoretical significance andrealistic value. While the diversity of information, complexity of social network and heterogeneityof individuals make this research challenging. Recently, the development of complex networktheory and computational experiments provides a new perspective to this research.Firstly, two typical information diffusion scenarios in social network are analyzed tosummarize the general rules of information diffusion in social network, based on whichclassification of information is raised. Then, agent based modeling is utilized to model the diffusionof advertisement and word-of-mouth in social network, which are two typical information diffusionexamples in social network. Fuzzy logic is introduced in agents’ decision methods to model theprocesses of recognition, persuasion, decision and diffusion of information, which are inspired bythe well-developed innovation diffusion theory. The bass modified model is constructed tounderstand the effect of word-of mouth on the diffusion of new products, based on whichword-of-mouth diffusing agents’ decision methods are modeled.Simulations and experiments of the agent based diffusion model of advertisement show thatmore influential initial agents can improve the diffusion rate, and economic incentive, as well asexternal environment can also affect the diffusion results. Simulations and experiments of the agentbased diffusion model of word of mouth show that word-of-mouth is most influential in the middlestage of the new product diffusion process, during which word-of-mouth marketing is mosteffective, and scale-free property is discovered in the word-of-mouth diffusion process.