Font Size: a A A

Research On The Social Network Centrality Node Identification Method Based On Evidence Theory

Posted on:2017-04-19Degree:MasterType:Thesis
Country:ChinaCandidate:M Y MaFull Text:PDF
GTID:2308330482992243Subject:Computer software and theory
Abstract/Summary:PDF Full Text Request
The stable relations system due to the social interactions between individuals is called social network. With the rapid development of the internet, the communication between people is becoming more and more convenient and fast, and the scale of social network becomes larger and larger.The relationship graph produced by social network also becomes larger. In recent years, because of the huge market potential and application value, social networks has get more and more researchers’ attention.The roles of all the elements in the social network are different, some elements are called central elements, which play important role in the connection of elements or the dissemination of information. The relationship graph produced by social network also becomes larger. Network centrality can reflect the power and status of individuals or organizations in social networks, the centrality of the study is the focus of social network research, which has great significance in mining the direction of propagation of information and the communication effect in the network. For example: sometimes there is a wide spread of false statements in the social network,by digging the central characters in the network, we can stop the spread of rumors in a timely manner.Although the node centrality research has very important meaning to the study of social networks, existing methods of identifying network centrality nodes are based on a central metrics. Each index has its advantages and disadvantages, which be used alone has certain one-sidedness. And the centrality of the node is relative, which is meaningful in the sub network of the node. So when we judging the centrality of the node, it is very necessary to divide the network into sub networks.Based on the existing study of node centrality,we summarized the concept and calculation method of all the centrality index comprehensively, and proposed the social network centrality node identification method based on the evidence theory. In this paper, all the index are considered. First of all, we get the indexes which closelyrelated to the node centrality through the experiment. And then model the social network which is to be calculated, calculate the shortest distance between nodes. By the shortest distance matrix, we carried out on the network hierarchy clustering and got the sub networks of the social network. Then we identified centrality nodes in one of those sub networks. The main process of the identification of centrality nodes as shown below. We separated the relationship matrix of the sub network from the social network, calculated the relevant indexes, processed relevant evidence and conflict evidence, assign the important degree of evidence. After these processing, evidence theory can be used for synthesis.By this method, we can analyze the centrality of nodes in the network comprehensively. And the recognition accuracy of the centrality node is higher, the recognition ability is stronger. We got good result in the Enron email data set.Therefore, the social network centrality node identification method based on the evidence theory has a strong research value and research significance.
Keywords/Search Tags:Social Network, Evidence Theory, Node Centrality, Sub Network Division, Enron Email
PDF Full Text Request
Related items