New paradigms in information visualization (poster session)

Abstract

We present three new wsualizatzon front-ends that aid navigatwn through the set of documents returned by a search engine (hit documents). We cluster the hzt documents to visually group these documents and label the groups with related words. The different front-ends cater for different user needs, but all can browse cluster information as well as drilling up or down in one or more clusters and refining the search using one or more of the suggested related keywords.documents, ie, when the query is refined to "computer hardware", a new set of related words emerges; for details see [9]. The obtained clusters based on this representation have proven to be meaningful in experiments with a large database of TREC human relevance judgements [12].The recent decade has seen much interest in information visualization, see eg [7,4,11,5,1,2]. In the following we suggest three new paradigms in information visualization displaying clusters of documents.

Multidimensional Visualization Techniques for Microarray Data

“…We notice a dip in the expressions in the center of the display, which is due to the switch from the SV40T dimensions to the WT dimensions and is not a true pattern in the dataset. The technique has been complemented by dimension ordering approaches, where the dimension order is determined by the structure of the data or the inherent class separation [31], [36][37] and 3D extensions [38]. One of the major disadvantages of Radviz is the overlap of points that occurs not only when the records have identical values on the displayed dimensions, but also when the records are scaled.…”

Neural-Network Enhanced Visualization of High-Dimensional Data

“…The technique has been used in several data domains including biomedical and complemented with dimension ordering techniques, where the dimension order is determined by the structure of the data, or the inherent class separation (Leban et al, 2005;Au et al, 2000;Grinstein et al, 2001;Bertini et al, 2005). We include an example of the technique as applied to the Iris data set, displaying the four dimensions of the data on the circle, and colouring the records by the flower type (Figure 2).…”

A new visual search interface for web browsing

“…Au et al [2] introduced a system to visually display the distribution of search results by grouping these results into topics according to their keywords. But their system does not allow progressively examining information in the search results as explored and supported in our system.…”