Vos: A New Method for Visualizing Similarities between Objects

Abstract

We present a new method for visualizing similarities between objects. The method is called VOS, which is an abbreviation for visualization of similarities. The aim of VOS is to provide a low-dimensional visualization in which objects are located in such a way that the distance between any pair of objects reflects their similarity as accurately as possible. Because the standard approach to visualizing similarities between objects is to apply multidimensional scaling, we pay special attention to the relationship between VOS and multidimensional scaling.

Keywords: Multidimensional Scaling, VOS, Horseshoe Effect, Sammon Mapping, Visualization of Similarities