Research On Visualization Optimization Of Network Relationship Diagram

Network-related data exists in every aspect of modern human society and virtual network society.Interpersonal telephone communication and e-mails constitute the communication network,and the friendships in social networking sites such as Weibo form social networks,while the departure points and destination constitute the urban transportation networks.The most common and reliable method to express network data is node-linked graphs visualization technology,which can help users to understand data quickly and accurately,and discover the value hidden in the data.With the expansion of data size,nodes and edges may be occluded each other in node-linked graphs,which hinders human cognition of data.Therefore,the crux of visualizing network diagram is how to arrange it efficiently and reasonably within a limited space.Based on the in-depth study of visualization methods of network relations,a point placement algorithm based on force-directed,and a edge layout algorithm based on edge bundling are proposed respectively,aiming at the two visual elements of nodes and edges in node link diagrams.In the proposed point placement visualization algorithm,a force calculation method which is independent of the distance between nodes is introduced.It is considered that the force acting on the nodes varies uniformly with the increase of iteration times.Therefore,it is not necessary to calculate the distance between two nodes,which greatly simplifies the calculation of the force and improves the efficiency.Unlike the classical force-directed point placement algorithm,the algorithm is not to minimize the total energy of the system,so it will not fall into the stalemate of local energy minimization.It will approach the global optimum effect and enrich the force-directed point placement model.In the proposed algorithm of edge layout visualization,the concept of edge bundling quality is introduced to improve the quality of edge bundling,hoping that the error value before and after the edge bundling operation is minimized.First,sampling and processing the edges by discretization method,and abstracting the edges into point clouds.Then,fitting the neighborhood nodes to bundling curves by the least square method.Next,projecting the discrete points to the fitted curves,and form the movement of nodes.Repeating the above process would produce curved bundles.This method could improve the quality of the edge bundling effectively,display the important structure of network graphs clearly,and reduce distortion at the extreme.Finally,the experiment shows that the two algorithms proposed in this paper can get the ideal visual layout effect,and realize the original intention of the research on the visualization optimization method in this paper.