12.11.2017

VisCoDeR

A tool for Visually Comparing Dimensionality Reduction Algorithms

Rene Cutura • Stefan Holzer • Michaël Aupetit • Michael Sedlmair

Abstract

We propose VisCoDeR, a tool that leverages comparative visualization to support learning and analyzing different dimensionality reduction (DR) methods. VisCoDeR fosters two modes. The Discover mode allows to qualitatively compare several DR results by juxtaposing and linking the resulting scatterplots. The Explore mode allows for analyzing hundreds of differently parameterized DR results in a quantitative way. We present case studies that show that our approach helps to understand similarities and differences between DR algorithms.

Video

High-res images

Screenshot of the discover mode Screenshots of the Discover mode — (a): linked textual description; (b): juxtaposed DR results, (i) component plane in the background, (ii) proximity view in the background, (iii) ray crossed by an other ray; (c): distribution of all data (grey), selected point and data of its class (pink) along input dimensions. Screenshot of the explore mode Screenshot of the Explore mode — (a): color legends of DR algorithms and their parameterizations; (b): control over the meta t-SNE; (c): direct encoded meta view with 1004 DR results; (d): clicking a dot in the meta view displays the DR result in detail. Screenshot of the discover mode; Juxtaposition of DR results of the MNIST dataset Screenshot of the juxtaposition of the DR results of the MNIST dataset — For the MNIST dataset, t-SNE offers the most correct result regarding labels as ground truth. The different digits are clearly separated based on their image pattern, while all DR ignore their class labels. Outlying digits can also be explained by referring to their snippet images showing the original digit data. Screenshot of the discover mode; Three types of overview Screenshot of three types of overview — (left): parallel coordinates; (middle): scattermatrix combined with density plots; (right): direct 3d visualization.

Downloads

Datasets

Spotify Iris MNIST S-shape Swissroll Waves

DR algorithms

PCA MDS LLE ISOMAP t-SNE

Tool

Link to VisCoDeR

Authors

René Cutura
rene.cutura@gmail.com
Stefan Holzer
st.holzer@gmx.net
Michaël Aupetit
maupetit@hbku.edu.qa
Michael Sedlmair
michael.sedlmair@univie.ac.at