Icon and Geometric Data Visualization with a Self-Organizing Map Grid

Autor(es):Mostrar menos -

Morais, Alessandra Marli M. ; Quiles, Marcos Goncalves ; Santos, Rafael D. C. ; Murgante, B ; Misra, S ; Rocha, AMAC ; Torre, C ; Rocha, JG ; Falcao, MI ; Taniar, D ; Apduhan, BO ; Gervasi, O

• Tipo de documento: Artigo Científico
• Fonte: COMPUTATIONAL SCIENCE AND ITS APPLICATIONS, PART VI - ICCSA 2014; v. 8584, p. 14-pg., 2014-01-01.

Resumo

Data Visualization is an important tool for tasks related to Knowledge Discovery in Databases (KDD). Often the data to be visualized is complex, have multiple dimensions or features and consists of many individual data points, making visualization with traditional icon- and pixel-based and geometric techniques difficult. In this paper we propose a combination of icon-based and geometric-based visualization techniques backed up by a Self-Organizing Map, which allows dimensionality reduction and topology preservation. The technique is applied to some datasets of simple and intermediate complexity, and the results shows that it is possible to reduce clutter and facilitate identification of associations, clusters and outliers. (AU)

• Processo FAPESP: 14/05453-6 - The 14th International Conference on Computational Science and Its Applications
• Beneficiário: Rafael Duarte Coelho dos Santos
• Modalidade de apoio: Auxílio à Pesquisa - Reunião - Exterior