Research On The Analysis Method Of Person Relationship For News Media

News media is one of the important objects for people to obtain information on the Internet.Analyzing the influence of figures in the news media network is helpful for social science researchers or economic and political researchers to understand the dynamics of international or social events,and to analyze the transmission power of information in the news.In order to deal with the complex information in the news conveniently and efficiently,the problems are often transformed into complex network-related problems for solving.In the existing researches on the influence of complex network nodes,only the local or global information of nodes is used,and the contribution of node influence and the connection between local and global information are ignored,which leads to poor effect of influence assessment.In addition,the existing methods only deal with homogeneous networks,but are not applicable to heterogeneous networks containing more node types and edge types.In view of the above problems,this paper uses the idea of complex network for reference,transforms the problem of evaluating the influence of news media figures into the problem of heterogeneous network analysis,and puts forward a method of analyzing the relationship between people in news media.According to the influence contributions of the neighbors of the nodes,the local importance is further excavated,and the comprehensive influence of the nodes is evaluated organically with the overall network structure information.The main work of this thesis is summarized as follows:(1)According to the relationship among the four entities,this thesis proposes a method to construct the heterogeneous network of the entities related to news.This method introduces four types of nodes corresponding to people,events,journalists and news,and excavates the possible information transmission relations among these nodes from news data,thus constructing a heterogeneous network reflecting the relations and structures among news entities.It is the basis for the influence assessment of character nodes.(2)On the basis of heterogeneous network construction,a measure of node influence,information entropy centrality,is proposed.The centrality measure refers to Shannon's idea of information entropy.The information transfer between nodes is abstracted as a link edge,and the information transfer from one node to another through the network path is regarded as a single independent event.The occurrence of an event will carry a certain amount of self-information,and the expectation of self-information is used as the information entropy centrality of the node.The centrality measure is used to evaluate the influence of character nodes and good results are obtained.(3)In order to further evaluate the node influence of people,a node influence assessment method based on importance contribution was proposed.Based on the centrality of information entropy and the contribution of importance of nodes,the method further mines the local information of nodes,and constructs the probability transition matrix of heterogeneous networks.The random transition algorithm of heterogeneous networks is used to mine the connection relations of network nodes to enrich the matrix information.Finally,the matrix iterative calculation is carried out.The experimental results show that the node influence evaluation method based on importance contribution has a good effect on the node influence evaluation of people.