Research On Social Network User Influence Analysis And Information Dissemination Modeling

With the popularity of the Internet and the wide application of web2.0 technology,the online social networks had been developed rapidly.At the same time,these social platforms have also changed people's social ways and daily life.Therefore,the construction and analysis of the social networks have attracted the attention of a large number of researchers.Social network influence analysis and information dissemination modeling are one of the important research contents of social network analysis,who are closely related and complementary to each other.The analysis of the social network influence can reveal the key factors of information dissemination.The study of information dissemination model can represent the mechanism of information transmission,which is the basis for the impact assessment of social networks.In view of this,this paper mainly analyzes the key nodes of social networks from the aspects of influence measurement criteria and implementation algorithms,and studies the information transmission modeling from the transmission and control of emotions in the process of information transmission.The main work and innovation points are summarized as follows:1.For mobile social networks,a centrality measurement method based on multi-social attributes weighted is proposed to provide a basis for mobile social networks to find important nodes.After the contact duration time being introduced into the time evolution model,we quantify various human social relations and mobile features to edge weight and a correction factor is introduced into the definition of tie strength(or weight).The newly quantized tie strength is denoted as the multi-social attributes weighted.Then the traditional centrality degree is extended with multi-social attributes weighted.Two coordination parameters are constructed,which are applied to the multi-social attribute degree centrality and shortest path,respectively.Furthermore,the multi-attribute shortest path is applied to closeness and betweenness centrality.The experimental results show that,compared with the social relation index,the multi-social attribute weighted can better represent the tightness between nodes.Multi-social attribute centrality can effectively identify important nodes in mobile social networks.There is a correlation between the multi-social attributes degree centrality and the multi-social attributes betweenness(or closeness)centrality.2.For dynamic social networks,an effective influence maximization algorithm is proposed to solve the dynamic influence maximization problem by updating activation probability of nodes which are influenced by newly added seed nodes.Under the dynamic independent cascade model,the monotonic sub-model property of the influence propagation function is proved theoretically,and the recursive formula is given.At the same time,the influence of the seed set obtained by this algorithm is at least approximately(1-1/e)of the optimal solution.The experimental results of three real social networks show that,no matter which probability model is,this method is superior to the two existing influence recognition methods in performance.It is also demonstrated by the experiment that when there are multiple rings in the social network,the actual influence is less than the calculated influence transmission.This particular condition that appears in the dynamic independent cascade model is similar to that in the SI model.The number of common seeds calculated by the former t network and t-1 network increases with the increase of the similarity of the two successive snapshot graphs of t and t-1.3.Aiming at dynamic multi-social networks with common users,a communication model of dynamic multi-social networks is established.In this model,multiple dynamic networks are merged into a dynamic network,in which the self-propagating edges of common users are added to the snapshots of each frame of the integrated network.Then the dynamic influence maximization algorithm in chapter 4 is extended to dynamic multi-social networks.Experimental analysis shows that the proposed model can not only accurately and vividly represent the dynamic characteristics,but also reflect the mutual influence of common users on multiple social networks.If the common users are chosen the nodes with greater influence in each network,the communication range of the integrated network is obviously larger than that of a single network,and the interaction of multiple dynamic social networks is more obvious.4.Aiming at a controversial hot event in social networks,a model of emotion transmission is proposed to provide theoretical basis for information transmission and emotion control in real networks.By analyzing the mechanism of emotion transmission in social networks,the differential equation of mean field theory transmission dynamics for each state is established,that is,the SE2 IR system model based on emotion.Aiming at the problem of maximizing positive emotions in social networks with cost constraints of control strategies,based on the optimal control theory,the control means and optimal control problem were proposed,and the optimal solution of controllable SE2 IR model was solved.For 16 kinds of combinations of four control means,using the optimal control system,we give the corresponding optimal control strategy in each case.For each optimal control strategy,the simulations contrast corresponding positive communication node density curve and control the cost.According to the price increase speed,positive communication nodes and the peak demand,it can rovide real emotional control to choose appropriate control strategy.Simulation results also show that the optimal control strategy under the condition of means 4 and its hybrid means can make the speed of positive propagation node faster and the peak value higher.Therefore,the optimal combination of means can be chosen according to the control cost.