Authors
Joseph A. Cottam
Ben Martin
Chris Mueller
Andrew Lumsdaine
Abstract
Visual Similiarity Matrices are an approach to representing the structure of large graphs. Different node ordering strategies have been shown to generate characteristic patterns in the resulting dotplot. This paper studies the effect of graph properties on the characteristic pattern of a breadth-first search ordering. The goal is to derive useful, quantitative information about the graph from the VSM. We find that several parameters of graph theoretic revelance can be inferred from the characteristics of the visual similiarity matrix image.
Index Terms
I.3.6 [Computer Graphics:]: Methodology and Techniques—Interaction Techniques G.4 [Mathematical Software:]: User Interfaces— [G.2.2]: Graph Theory— [H.5]: InformationSystems—Information Interfaces and Presentation