Research On Social Network Information Propagation Model Based On SIR Model

In recent years,the rapid spread of mobile Internet promotes the social networks to be an increasingly important part in people's daily lives.It has been preferred choice for people to publish information and acquire information from social networks.Among them,more and more commercial promotions are performed by means of social network for business information dissemination.In the thesis,the dissemination process of preferential information published by business companies in Microblog network are studied,and the dynamic model of preferential information dissemination are intensively analyzed.Contributions of the thesis can be summarized as follows:?1?Based on the SIR?susceptible-infected-recovered?model,a SEAIR?susceptible-lurker-super-normal-recovered?dynamic model is established.In the SEAIR model,three basic compartment status of super-spreaders,normal-spreaders and lurkers are introduced.Moreover,each state increases the probability of entering and leaving,enhancing the dynamics of the SEAIR model.Super-spreaders and normal-spreaders have different intensity of infection parameters?₁ and?₂ in the information dissemination process,and the lurkers are used to describe the hesitation process of the information dissemination process.In addition,the dynamic equation of the SEAIR model is established based on the mean field theory,and the basic infection number?ie threshold??₀ is derived.The theory proves that when?₀ is less than 1,the SEAIR model has local and global stability in the state of free-information equilibrium,and the propagation process tends to disappear.When?₀ is greater than 1,the SEAIR model has a unique local equilibrium state.?2?In order to verify the performance of the SEAIR model,the Baidu App promotion activity in Weibo was taken as an example to numerically simulate the SEAIR model and the SIR model.The simulation results show that compared with the SIR model,the root mean square error?RMSE?of the SEAIR model is reduced by33.2%,and the average absolute error of the SEAIR model is reduced by 24.74%.At the same time,we evaluate the influence of key parameters?₂,?,?₂,?₁,?,?₁,?₃,u on the performance of the SEAIR model.The analysis of the infection rate?₂,conversion ratio?and degradation ratio?₂ parameters of the super-spreaders shows that the new compartment A state?super-spreaders state?can greatly expand the scope and speed of information dissemination by retweet,making the information highly influential in the early stage.The information that the super-spreaders is concerned with will not only peak in advance,but also have a longer influence.The characteristic analysis of the normal-spreader's infection ratio?₁,conversion ratio?and degradation ratio?₁parameters shows that the influence of normal-spreaders is much lower than that of super-spreaders.The conversion ratio?of the normal-spreaders determines the number of users who choose to transmit,and plays a decisive role in information dissemination in the real network.The degradation rate?₁ of the normal-spreaders greatly affects the propagation peak of the information,and it has a great influence on the continuous propagation and forwarding state of the information.The lurker E and the entry and leaving parameters u characterize the dynamic change characteristics and hesitation state during information propagation.?3?Based on the SEAIR model,an analysis method for the trend of advertising information dissemination on Weibo is proposed in this paper.The method is mainly divided into four parts:data collection,construction of spatial network,microblog advertisement information dissemination dynamics analysis and output of analysis result.The method first obtains the relevant information of the advertisement through the data collection method;then constructs the corresponding network by the method suitable for constructing the spatial network structure;And use the SEAIR model to describe the corresponding network,and determines the parameter value of the model in the propagation process;Finally,the initial information of the advertisement information is brought into the model for determining the parameters,and the dynamic change of the whole advertisement information propagation process under the method is obtained.