Sanity check for class-coloring-based evaluation of dimension reduction techniques

Research output: Chapter in Book/Report/Conference proceedingConference contribution

8 Citations (Scopus)

Abstract

Dimension Reduction techniques used to visualize multidimensional data provide a scatterplot spatialization of data similarities. A widespread way to evaluate the quality of such DR techniques is to use labeled data as a ground truth and to call the reader as a witness to qualify the visualization by looking at classcluster correlations within the scatterplot. We expose the pitfalls of this evaluation process and we propose a principled solution to guide researchers to improve the way they use this visual evaluation of DR techniques.

Original languageEnglish
Title of host publicationACM International Conference Proceeding Series
PublisherAssociation for Computing Machinery
Pages134-141
Number of pages8
Volume10-November-2015
ISBN (Print)9781450332095
DOIs
Publication statusPublished - 10 Nov 2014
Event5th Workshop on Beyond Time and Errors: Novel Evaluation Methods for Visualization, BELIV 2014 - Paris, France
Duration: 10 Nov 2014 → …

Other

Other5th Workshop on Beyond Time and Errors: Novel Evaluation Methods for Visualization, BELIV 2014
CountryFrance
CityParis
Period10/11/14 → …

    Fingerprint

Keywords

  • Dimension reduction
  • Evaluation
  • Scatter plots

ASJC Scopus subject areas

  • Human-Computer Interaction
  • Computer Networks and Communications
  • Computer Vision and Pattern Recognition
  • Software

Cite this

Aupetit, M. (2014). Sanity check for class-coloring-based evaluation of dimension reduction techniques. In ACM International Conference Proceeding Series (Vol. 10-November-2015, pp. 134-141). Association for Computing Machinery. https://doi.org/10.1145/2669557.2669578