Sets Visualization using their Graph Representation

Abstract

With the emergence of new data acquisition technologies, large amounts of data are available in many domains. While a significant amount of computational research is dedicated to the analysis of such data, it is needed to be visualized in a way that is easy to analyze and understand. Recently, there have been significant advances in visualizing graphs; however, not enough tools exist for automatic visualization of sets. In this paper, we devise a spectral approach for visualizing overlapping sets, so that the underlying hierarchy and relations of the sets can be easily understood by visual inspection. The algorithm utilizes the spectral decomposition of the graph representation of the sets to compute the best coordinates for all items on the Euclidean plane. The experimental results were very encouraging, and showed positive indication on the efficiency of the proposed method.