Interactive Bivariate Mode Trees for Visual Structure Analysis
Abstract
The number of modes in a kernel density estimation of a certaindata distribution is strongly dependent on the chosen scale parameter.In this paper, we present an interactive mode tree visualizationthat allows to visually analyze the modality structure of a datadistribution. Due to the branched structure of the bivariate modetree, composed of many curved arcs in 3D, we need to utilize advancedtechniques, including clutter removal through transparency,on demand outlier suppression or preservation, and best views, toimprove the legibility of the visualization mapping.
M. Florek and H. Hauser, "Interactive Bivariate Mode Trees for Visual Structure Analysis," in Proceedings of the Spring Conference on Computer Graphics (SCCG 2011), 2011, p. ??–??.
[BibTeX]
The number of modes in a kernel density estimation of a certaindata distribution is strongly dependent on the chosen scale parameter.In this paper, we present an interactive mode tree visualizationthat allows to visually analyze the modality structure of a datadistribution. Due to the branched structure of the bivariate modetree, composed of many curved arcs in 3D, we need to utilize advancedtechniques, including clutter removal through transparency,on demand outlier suppression or preservation, and best views, toimprove the legibility of the visualization mapping.
@INPROCEEDINGS {florekHauser11modeTree,
author = "Martin Florek and Helwig Hauser",
title = "Interactive Bivariate Mode Trees for Visual Structure Analysis",
booktitle = "Proceedings of the Spring Conference on Computer Graphics (SCCG 2011)",
year = "2011",
pages = "??--??",
abstract = "The number of modes in a kernel density estimation of a certaindata distribution is strongly dependent on the chosen scale parameter.In this paper, we present an interactive mode tree visualizationthat allows to visually analyze the modality structure of a datadistribution. Due to the branched structure of the bivariate modetree, composed of many curved arcs in 3D, we need to utilize advancedtechniques, including clutter removal through transparency,on demand outlier suppression or preservation, and best views, toimprove the legibility of the visualization mapping.",
images = "images/florekHauser11modeTree.jpg, images/florekHauser11modeTree2.jpg",
thumbnails = "images/florekHauser11modeTree_thumb.jpg, images/florekHauser11modeTree2_thumb.jpg",
location = "Budmerice, Slovakia"
}