Research On Information Tracking And Propagation Scheme Based On Social Influence

With the rapid development of the Internet,social platform,bringing unprecedented convenience to human life,has become one of the indispensable platforms for people to exchange,publish and acquire information.Meanwhile,due to the high degree openness and freedom of social platforms,user behavior becomes extremely complicated,which directly leads to the process of online information dissemination becoming complicated and difficult to control,and brings certain threats to the network security and social stability.It is an urgent problem that how to find out the law of information dissemination,explore the key elements in the process of dissemination,and establish effective preventive measures in the social network.This thesis attempts to start with the online social network and study on the law of information dissemination in the network.The main work and contributions are as follows:1.In view of the three key elements in the hotspot topic network,this thesis proposes a social hotspot tracking scheme named MPURank based on the tripartite graph and multi-messages iterative driven.First,the message propagation network tree model is introduced,and the network topology formed by the user retweeting is used to determine the propagation path of the forwarding user,and the message propagation tree is reconstructed.Second,the topic network is regarded as a network of multi-message dissemination simultaneously and the multi-message topic propagation tree network is reconstructed based on the single-message propagation tree model,and the topic tripartite graph is established.Especially,it can also establish a correlative relationship among the three key elements by using tripartite graph,using a related idea to identify the three key elements rather than an isolated ranking.Finally,based on the sorting algorithm HITS,an iterative scoring algorithm is proposed,which is used on the topic tripartite model to obtain the key score sequences of elements.2.Aiming at the mechanism of multi-message propagation of hot topics in the network,an improved propagation dynamics model is proposed.First,the interaction ability of the participating nodes is obtained from the internal and external factors,which uses the characteristic attributes and behavior information.Furthermore,the average state transition probability of the node is calculated by using the node binomial distribution principle,which is used to replace the infection rate in the kinetic model,and the recovery rate is also obtained.At the end,combined with the average field theory,the trend of multi-message propagation network is simulated based on the kinetic model,and the interaction mechanism is researched in the multi-message propagation process under the same topic.Experimental verification was performed using real datasets of Sina Microblog based on the above model.The experimental results show that the model and optimized algorithm can mine many key elements in multi-message propagation networks in this thesis.Meanwhile,this method can also simulate the trend of multi-message propagation and further explore the interaction relationship in the process of multi-message communication.