The Community for Technology Leaders
RSS Icon
Subscribe
Issue No.06 - November/December (2003 vol.15)
pp: 1460-1471
Li Yang , IEEE
ABSTRACT
<p><b>Abstract</b>—This paper discusses 3D visualization and interactive exploration of large relational data sets through the integration of several well-chosen multidimensional data visualization techniques and for the purpose of visual data mining and exploratory data analysis. The basic idea is to combine the techniques of grand tour, direct volume rendering, and data aggregation in databases to deal with both the high dimensionality of data and a large number of relational records. Each technique has been enhanced or modified for this application. Specifically, positions of data clusters are used to decide the path of a grand tour. This cluster-guided tour makes intercluster-distance-preserving projections in which data clusters are displayed as separate as possible. A tetrahedral mapping method applied to cluster centroids helps in choosing interesting cluster-guided projections. Multidimensional footprint splatting is used to directly render large relational data sets. This approach abandons the rendering techniques that enhance 3D realism and focuses on how to efficiently produce real-time explanatory images that give comprehensive insights into global features such as data clusters and holes. Examples are given where the techniques are applied to large (more than a million records) relational data sets.</p>
INDEX TERMS
Data clustering, footprint splatting, grand tour, high-dimensional data, visual data exploration, volume rendering.
CITATION
Li Yang, "Visual Exploration of Large Relational Data Sets through 3D Projections and Footprint Splatting", IEEE Transactions on Knowledge & Data Engineering, vol.15, no. 6, pp. 1460-1471, November/December 2003, doi:10.1109/TKDE.2003.1245285
16 ms
(Ver 2.0)

Marketing Automation Platform Marketing Automation Tool