CIRCULARVIZ: AN ENHANCED RADVIZ VISUALIZING FOR MULTIDIMENSIONAL DATA

Abstract

Visualizing for multi-dimensional data points in data space is a central topic with many supplications in many branches of science. RadViz visualizes multi-dimensional data points in the 2D visual space based on force-based methods. Radviz visualizes information about relationships of multi-dimensional data points of data space. The traditional Radviz method finds the best order of anchor points on the circumference of the circle with radius one to enhance the relationships of multi-dimensional data sets. In our paper, we modify the Radviz method named CircularViz method that improves the Radviz layout for the group’s preservation the relationships of multi-dimensional data sets. The idea of the CircularViz method presents data dimensions by circular segments on the circle with radius 1 and finds the most suitable to gaining insights into class structures of multi-dimensional data sets. Our approach make provision an augmentation in visualizing group structures of multi-dimensional data on the 2D visual space. The effectiveness of the CircularViz method evidences by quality visualization measurements and several supervised high-dimensional data sets.