Visualizing Dimensionality Reduction Artifacts: An Evaluation - Inria - Institut national de recherche en sciences et technologies du numérique Accéder directement au contenu
Pré-Publication, Document De Travail Année : 2017

Visualizing Dimensionality Reduction Artifacts: An Evaluation

Résumé

Multidimensional scaling allows visualizing high-dimensional data as 2D maps with the premise that insights in 2D reveal valid information in high-dimensions, but the resulting projections always suffer from artifacts such as false neighborhoods and tears. These artifacts can be revealed by interactively coloring the projection according to the original dissimilarities relative to a reference item. However, it is not clear if conveying these dissimilarities using color and displaying only local information really helps to overcome the projections artifacts. We conducted a controlled experiment to investigate the relevance of this interactive technique using several datasets. We compared the bare projection with the interactive coloring of the original dissimilarities on different visual analysis tasks involving outliers and clusters. Results indicate that the interactive coloring is effective for local tasks as it is robust to projection artifacts whereas using the bare projection alone is error prone.
Fichier principal
Vignette du fichier
evalproj.pdf (2.8 Mo) Télécharger le fichier
Origine : Fichiers produits par l'(les) auteur(s)
Loading...

Dates et versions

hal-01523028 , version 1 (16-05-2017)

Identifiants

Citer

Nicolas Heulot, Jean-Daniel Fekete, Michael Aupetit. Visualizing Dimensionality Reduction Artifacts: An Evaluation. 2017. ⟨hal-01523028⟩
160 Consultations
185 Téléchargements

Altmetric

Partager

Gmail Facebook X LinkedIn More