Reduction of Complexity: An Aspect of Network Visualization

摘要

Networks are topological structures composed of nodes and arcs. Often, networks are visualized by point symbols and lines illustrating the nodes and the arcs, respectively. As the number of the nodes and links increases, the visual representation of the network needs generalization in order to keep the visual clarity of the image. To solve the problem considered, a methodology to aggregate nodes and links into hypernodes and hyper-links is developed. The algorithm, which is based on an information theoretic approach to reorder the adjacency matrix of the network, can generate hierarchies of hyper-networks. This kind of generalization algorithm can be used to construct images which visualize the main structures of the network. Some case studies demonstrate the algorithm.