Research On Visualization Of Massive Graph

Graph visualization is the sub-area of information visualization and the hot research direction.By transformed into graph and presented with graph visualization technology,the entity-relations information can be understood easily.With the "big data" era coming,the rapid expansion of entities from real world led to a surge in the amount of graph.Faced the graph containing tens of thousands of nodes and edges,the traditional algorithm is powerless in the practical application,the layout result also become chaos.Based on Force-Directed Algorithm(FDA)and multilevel approach,we propose a drawing algorithm aiming at massive graph.The algorithm can efficiently avoid the problem that the traditional algorithm has poor local minimum and slow running speed because of huge data amount.Our main contributions are summarized as blow.First of all,we propose two coarsening algorithm based on Weighted Edge Matching(WEM)and Weighted Maximal Independent Vertex Set(WMIVS).The WEM coarsening algorithm is mild and the WMIVS coarsening is more aggressive but universe.Both the two coarsening algorithm can preserve the cluster structure of massive graph.Furthermore,we propose BH-FR algorithm which combines the classic FR algorithm and Barnes-Hut model.On the basis,combining WEM and WMIVS coarsening algorithm,we propose an algorithm named Massive Graph Layout with Multilevel Approach(MGLMA).The algorithm is more in line with the accepted aesthetic standards,yet improves the layout efficiency of massive graphs.Finally,according to the research of this paper and visual analysis of entity-relations information,we designed and implemented a multi-dimensional platform for entity-relations information visualization based on graph.The system provides users with a variety of analytical methods and visualization tools.