Dynamics Of Explosive Information Spreading On Complex Network

Many communication phenomena in real life can be abstracted as information spreading on complex networks,such as the spreading of news,the propagation of rumors,the promotion of new products and the contagion of panic.Using complex network theory to study the mechanism of information spreading in real world has important practical significance.For example,understanding the rules of rumors spreading will help the government department concerned to make some early warnings and prevention strategies,which can mitigate or eliminate the harm to economic and social stability.Meanwhile,mastering the promotion mechanism of new products in social networks will help the merchants to predict the penetration rate of products and design a better promotion strategy.Thanks to the rapid development of network science,the problems of information spreading on complex networks have been extensively studied over the past 20 years,and a great deal of significant progress has been obtained,especially in terms of spreading threshold,contagion pattern and propagation paths.In recent years,considering the differences between information and epidemic spreading,many key factors of information spreading have also been unveiled,such as memory effects,social reinforcement and non-redundant contacts.However,most of these studies are based on phenomenological models and lacked the supporting of real data.Moreover,the advances in information technology and the popularity of various social services have substantially changed our life.People can learn about current events through the Internet,and they can also communicate online with friends on the other side of the earth.However,it is still not clear how these new features of communication activities affect the information spreading process,and it is still lacking the relevant theoretical analysis.Thus,in this thesis,we collected the information spreading data of many typical events on Sina Weibo,and discovered the explosive phenomenon of information spreading from the data: some attractive information can be accepted by many people overnight.By constructing new spreading models and developing relevant theoretical analysis methods,we show that besides the unique attributes of information content,the new features of communication activities through multiple channels play a critical role in boosting explosive information spreading.Specifically,we summarize the main contributions of this thesis as following:(1)By focusing on the information spreading data of six typical events on Sina Weibo,we surprisingly find that the spreading of modern information shows some new features depending on the individual events,i.e.,some information is spread extremely fast and outbreak overnight,while others are slowly and gradually out of our sights.To understand its mechanism,we present a Susceptible-Accepted-Recovered model with both information sensitivity and social reinforcement.Numerical simulations show that the model can reproduce the main spreading patterns of the six typical events.By this model,we further reveal that the spreading can be speeded up by increasing either the strength of information sensitivity or social reinforcement.Depending on the transmission probability and information sensitivity,the final accepted size can change from continuous to discontinuous transition when the strength of the social reinforcement is large.Moreover,we present an edge-based compartmental theory to explain the numerical results.These findings not only deepen our understanding of information spreading mechanisms,but also provide a reference for controlling explosive information spreading.(2)Information spreading has been studied for decades,but the reason why the information can be accepted by a large number of people overnight has not been understood well.Meanwhile,it is not clear whether the reason caused by the new features of multi-channels communications.In this study,based on a Susceptible-AcceptedRecovered(SAR)model,we examine the effects of two channels represented by a twolayered network,in which one channel is the intra-links within the same layer and the other is interlinks between layers.Different with the case of one single channel on a onelayered network,in the case of two channels,the spreading can be speeded up by the increased of the coupling strength,i.e.,average degree and transmission probability between the two layers.Strikingly,even if the social reinforcement is small,the strongly coupling strength can induce explosive transition in the information spreading process.Additionally,a big gap closed to the critical point for the explosive transition was found in the phase space of theoretical analysis,which indicates the emergency of a global largescope outbreak.These findings may be of great significance on the understanding and controlling explosive information spreading in modern society.(3)A common point in many of the past studies is that there is only one transition in the phase diagram of the final accepted size versus the transmission probability.However,whether other factors will alter this phenomenology is still under debate,especially for the case of information spreading through many channels and platforms.In the present study,we adopt a two-layered network to represent the interactions of multiple channels and propose a Susceptible-Accepted-Recovered information spreading model.Interestingly,our model shows a novel double transition including a continuous transition and a following discontinuous transition as explosive spreading in the phase diagram,which originates from two outbreaks between the two layers.Furthermore,we reveal that the key factors are a weak coupling condition between the two layers,a large adoption threshold and the difference of the degree distributions between the two layers.Moreover,we also test the model in the coupled empirical social networks and find similar results as in the synthetic ones.Then,an edge-based compartmental theory is developed which fully explains all numerical results.Our findings may be of significance for understanding the secondary outbreaks of information in real life.The results of this thesis show that information sensitivity,social reinforcement,transmission probability and multi-channels communication methods play an important role in explosive information spreading.Uncovering more key factors and understanding in-depth the inherent laws of explosive information spreading are yet waiting for us to explore in the future.