A vector field visualization technique for self-organizing maps

  • Authors:
  • Georg Pölzlbauer;Andreas Rauber;Michael Dittenbach

  • Affiliations:
  • Department of Software Technology, Vienna University of Technology, Vienna, Austria;Department of Software Technology, Vienna University of Technology, Vienna, Austria;eCommerce Competence Center – ec3, Vienna, Austria

  • Venue:
  • PAKDD'05 Proceedings of the 9th Pacific-Asia conference on Advances in Knowledge Discovery and Data Mining
  • Year:
  • 2005

Quantified Score

Hi-index 0.00

Visualization

Abstract

The Self-Organizing Map is one of most prominent tools for the analysis and visualization of high-dimensional data. We propose a novel visualization technique for Self-Organizing Maps which can be displayed either as a vector field where arrows point to cluster centers, or as a plot that stresses cluster borders. A parameter is provided that allows for visualization of the cluster structure at different levels of detail. Furthermore, we present a number of experimental results using standard data mining benchmark data.