Research Of Large-scale Information Visualization Of Social Network Method

Visualization technology is one of the effective means of social network analysis.Visualization technology transforms network data information,network structure,and connections between nodes into graphical forms.Due to its intuitive,effective,and visual characteristics,users can gain a clear understanding.At the same time,in order to realize a deeper understanding and mining of social networks,users can further explore,analyze and study the local or key inter-node connections of the network through the processing of visual display effects and the corresponding interaction methods of users.However,with the vigorous development of Internet science and technology,the scale of network data is also expanding increasingly.The research content of social network information visualization technology also focuses on how to display large-scale network data intuitively,clearly and effectively.The thesis studies the basic concepts and characteristics of social networks,the layout methods and analyzes the existing problems of existing graph layout methods.The layout results of the force-directed for the mainstream graph layout algorithm often have visual confusion of node overlap and coverage,difficulty in distinguishing structural information,and lack of highlighting key nodes in the network.The paper improves the classical force guidance layout algorithm-FR algorithm.When calculating the force between nodes,considering the degree of the node and the size of the node itself,the "boundary-toboundary" distance between the nodes and the nodes that prevent overlap is introduced,and the repulsive force is adjusted.According to the degree of centrality and betweenness centricity of the nodes,the key nodes are highlighted,and the clustering is good,the community is clearly divided,and the visual layout results are clearly displayed by the key nodes.For large-scale social networks with geographical attributes,the layout of the vertices in the network graphs will destroy the original geographic information,and the layout from the side will lead to serious edge crossing problems.To this end,the paper proposes a layout method that combines map projection with force-guided edge bundling.Firstly,the Mercator projection method in map projection is used to determine the distribution of nodes,objectively reflecting the geographical characteristics of the network,and truly showing the specific geographical layout effect.Then,by using the force-directed edge bundling algorithm,the problem of visual confusion of edge crossing is effectively solved,and the distribution of community structure and nodes is clearly displayed,which is convenient for obtaining useful information.Based on Matlab,Eclipse,Netbeans and HTML,the paper designs and preliminarily implements a social network information visualization prototype system,including data reading,data cleaning,node centrality evaluation,network index calculation,visual data generation,and graph layout visualization.The system integrates the improved algorithms described above and provides a preliminary solution for large-scale social network visualization.The results show that the system designed by the thesis can fully display the community structure of the social network and complete the corresponding visualization tasks.