Multidimensional clusters in RadViz

Abstract

The paper reviews those properties of RadViz visualization method [2] mapping data from n dimensional space into the plane which are important for identification of clusters in the multidimensional original data. It uses 2 characteristic examples of datasets which clearly point to a certain drawback of the original RadViz mapping. The identified problem can be resolved using 2 minor modifications of the RadViz algorithm which are suggested in the last section of this contribution.