University of Houston
Department of Computer Science
In
partial fulfillment of the Requirements for the Degree of
Master of
Science
Charoenchai Sutippantupat
will defend his
thesis
Distance Preserving Mappings for Scientific Data
Visualization
Abstract
The features of a tool for exploratory data analysis are presented
that supports the visualization of n-dimensional objects using 2-dimensional
displays. The tool employs evolutionary computing, a hill climbing algorithm, and
a hybrid algorithm to find good distance preserving mappings
from n-dimensional space to two-dimensional space. Results of experiments
are discussed that compare and evaluate the three different approaches.
Visualization techniques are employed by the tool to provide a graphical representation
of two-dimensional data and to allow its user to
analyze data with helpful features such as zooming, annotation, and multiple
views. Results of case studies are described that used the tool for
analyzing a NBA-player dataset, a university dataset, and a geometrical dataset.
Moreover, we also compared our approach with multidimensional scaling, a popular
technique for visualizing n-dimensional data, using the three datasets. The
experimental results are encouraging in that the employed hybrid approach that
combines hill climbing with evolutionary computing techniques produced a smaller
error than multidimensional scaling for the three benchmarks.
Date: Tuesday, October 3, 2000
Time: 1:15 PM
Place:
550-PGH
Faculty, students, and the general public are
invited.
Thesis Advisor: Dr. Christoph F. Eick