Research On Word-of-mouth Dissemination Model In Online Social Networks

With the development of internet technology,online social networks have become the main channel for people to exchange and share information.Large amounts of word-of-mouth(WOM)information have emerged in online social networks,which has an important impact on consumers' purchasing decisions and brand reputation.The prediction and control of the diffusion of WOM has attracted attention from enterprises.The complexity of online social network structure and the diversity of consumer behaviors have brought great challenges to the analysis of WOM diffusion.Therefore,studying the mechanism and law of WOM diffusion has become a popular topic in academic research.By summarizing the limitations of current researches,this paper applies the theory and method of complex network and propagation dynamics to study the WOM communication model in online social networks from the perspective of WOM communication network,WOM itself and consumers.The main research objectives are as follows:(1)In view of the fact that online social networks often have an overlapping community structure,based on the infectious disease model,individual behaviors such as consumer selfishness are considered,and a WOM diffusion model in online social networks with overlapping-community structure is constructed.Community overlapping-level parameters are introduced to describe the extent of community overlapping.In addition,consumer selfishness parameter and social selfishness parameter are incorporated to describe the selfish behavior of consumers towards friends and strangers.The results of numerical experiments show that in social networks with higher overlapping-degree communities,WOM information spreads faster and reaches a wider range.Meanwhile,consumer individual selfishness and social selfishness significantly affect the speed and scope of WOM information dissemination in social networks,and the increased overlapping-degree of community structure can reduce the impact of selfishness.In addition,consumers may discard information after receiving WOM information.When the discarding probability increases to a certain extent,WOM information may disappear in the network without spreading,but the overlapping-community structure can mitigate the impact of discarding behavior.(2)Based on the existence of large amounts of WOM information in online social networks,and the phenomenon of information competition in the process of diffusion,a WOM diffusion model that considers information attraction and competition is constructed.A WOM attraction function is proposed to characterize the time-static information attraction of long-cycle products and time-varying information attraction of short-cycle products.Maximum attractiveness,information outbreak time,and information outbreak index is used to describe the characteristics of time-varying attractiveness.Numerical experiments show that the three aspects of attractiveness significantly affect the WOM information dissemination.Improving the attractiveness in the initial stage is important for enhancing information competitiveness.When WOM information with static attractiveness and that with time-varying attractiveness compete,if the maximum attractiveness of the former is lower than that of the latter,the diffusion scale of the former will not exceed that of the latter.However,the information diffusion of the former can be improved by increasing the maximum attractiveness or adjusting the outbreak time and outbreak index.(3)Considering the willingness to receive WOM information and herding behaviors,a WOM diffusion model based on herding behavior is constructed.Herding factor is incorporated in the model.Herding attitude towards friends and strangers are considered.Numerical analysis show that the willingness to receive WOM information and herding behaviors significantly influence the information diffusion.When consumers have low willingness to receive WOM information,a low level of herding behavior will promote the information diffusion.Meanwhile,at the early stage of the diffusion,high willingness to receive information or high level of herding behavior will suppress information diffusion.In addition,in networks without herding behavior,WOM information spread fast at the early stage,and in networks with herding behavior,WOM information diffuse slow at the early stage and then will suddenly outbreak,and will spread faster than that in networks without herding behavior.Based on the actual online review data,the effectiveness of the proposed word-of-mouth propagation model was verified in two steps.The proposed model is used to fit the actual data propagation process,and a good fitting effect is obtained.The influence of herd behavior on word-of-mouth communication is verified by topic mining method.In summary,this paper extends the theory and method in research of WOM diffusion in online social networks,which will help to understand the characteristics of WOM dissemination,and has crucial theoretical and practical significance for the prediction and control of WOM diffusion.