Analysis Of The Impact Of Information Propagation Based On Node Relation Evaluation In Social Networks

With the rapid development of Internet technology,social network platforms have become an important medium and carrier for modern information dissemination.The dissemination of network information has brought tremendous changes to people's lives and work,which made it possible for people to communicate anytime and anywhere.In the meanwhile,the dissemination of online public opinion has attracted more and more attention from enterprises and governments.Research and analysis of the laws of information dissemination can provide important control ideas and methods for government and corporate public opinion monitoring and other work,and it is worthy of multi-level and multi-dimensional in-depth the study.At the present stage,the research work on information dissemination is mainly focused on the prediction of information dissemination-related issues based on the specific observable network structure.However,relatively few studies have considered the use of link prediction principles to further dig out potential social network topologies for deeper analysis of information dissemination mechanisms and laws.At the same time,most of the existing research on information dissemination models are based on unsigned social network designs,only considering that there may be future links established during the evolution of the network structure,without considering the types of new links,but the type of links in the actual social network will have a direct impact on the user's behavior.Therefore,it is necessary to consider designing an information dissemination model suitable for sign social networks from the perspective of relations and considering the polarity of relations.In view of the deficiencies in the previous research work,this paper focuses on the two main issues involved in the analysis of the impact of information propagation based on node relation evaluation:1.How to use observable network structure to predict potential links and distinguish link types;2.On the basis of link prediction,how to model the information dissemination process by considering the positive relationship and negative relationship between nodes,and further quantify the analysis.In response to the above two issues,this paper proposes a social network information propagation model based on the implicit relationship mining framework and considering the existence of negative relationships.The main work is as follows:(1)In the research of implicit relationship and type prediction based on social networks,firstly,implicit relationship mining in the network is to predict the possible future links in the network by adding the intimacy index on the basis of the traditional link prediction algorithm.Then,on the basis of implicit relationship mining,we further judge the types of these relationships.Finally,we output the potential relationships in the network and the results of type prediction by integrating the first two steps.(2)In the research of social network information propagation model integrating negative relationships,firstly,mathematically model the influence of the relationship between nodes on the attitude of the nodes,and then model the update process of the attitudes of the nodes,and give the corresponding propagation rules.Finally,based on the modeling results obtained in the above two steps,the modeling of the node state change process in the entire communication process is completed,and an information communication model suitable for the fusion of negative relationships in symbolic social networks is obtained.In the meanwhile,the propagation characteristics can be further analyzed on the basis of the above-mentioned implicit relationship mining.(3)Experiments with the above models are conducted on the data from the real data sets of three online social networking sites.The experiments prove that the above models are better than the classic models in performance,and can get more accurate relationship and type prediction results and more reasonable information dissemination characteristics analysis results.