Share Email Print

Proceedings Paper

Geodesic self-organizing map
Author(s): Yingxin Wu; Masahiro Takatsuka
Format Member Price Non-Member Price
PDF $17.00 $21.00

Paper Abstract

Self-Organizing map (SOM) is a widely used tool to find clustering and also to visualize high dimensional data. Several spherical SOMs have been proposed to create a more accurate representation of the data by removing the “border effect”. In this paper, we compare several spherical lattices for the purpose of implementation of a SOM. We then introduce a 2D rectangular grid data structure for representing the geodesic dome. This new approach improves the neighborhood searching process in the spherical gird. The new Geodesic SOM and its data structure are tested using socio-demographic data. In the experiments, we try to create a notion of direction in the Geodesic SOM. The direction facilitates more consistent visual comparison of different datasets as well as to assist viewers building their mental maps.

Paper Details

Date Published: 11 March 2005
PDF: 10 pages
Proc. SPIE 5669, Visualization and Data Analysis 2005, (11 March 2005); doi: 10.1117/12.586807
Show Author Affiliations
Yingxin Wu, Univ. of Sydney (Australia)
Masahiro Takatsuka, Univ. of Sydney (Australia)

Published in SPIE Proceedings Vol. 5669:
Visualization and Data Analysis 2005
Robert F. Erbacher; Jonathan C. Roberts; Matti T. Grohn; Katy Borner, Editor(s)

© SPIE. Terms of Use
Back to Top