Radviz extensions with applications
Georges Grinstein, John Sharko · 2009
RadViz (RV), a visualization tool developed at the Institute for Visualization and Perception Research at the University of Massachusetts Lowell, has proved to be very useful in a wide variety of applications. It has also been incorporated internationally into several generalized visualization systems. This research represents efforts to extend the ability of RV to visualize complex datasets. RV has the particular capability of being able to effectively display high dimensional datasets. This research has made use of this property by developing Vectorized RadViz (VRV). VRV provides the capability to enhance the power of RV by creating dimensional anchors assigned to individual values coming from each of the dimensions in the dataset. This tends to significantly increase the number of dimensions to be displayed, again, harnessing RV's strength to display high dimensional datasets. Repositioning these dimensional anchors expands the ability of RV to expose underlying characteristics of the dataset. This research shows how VRV can be applied to cluster ensembles and decision trees and how RV can display fuzzy clusters. Using the extent to which each record is a member of each cluster of the fuzzy cluster set RV is able to display their relationships. Fuzzy clusters were also compared to cluster ensembles using VRV. The basic features of each of the visualizations in this research were illustrated using the Iris dataset but also applied to larger scale microarray datasets. Note: Much of this research has already been published in references (1) and (2).