University of South Florida
Using Hyper-Dimensional Spanning Trees to Improve Structure Preservation During Dimensionality Reduction
Abstract
dc:descriptionUnderstanding relations in hyper-dimensional data is a prevalent problem, which is often approached by using dimensionality reduction. The structure preserved from the original data is often dependent on the type of dimension reduction algorithm used, and it can produce results that vary substantially from one another. Visualizing hyper-dimensional data helps to understand the data, but it presents a problem, as our visualizations rely upon two or three-dimensional displays. Current dimension reduction methods, used to reduce hyper-dimensional data to low-dimensional data, often produce results that fail to preserve the structure as the complexity of the data increases. This reduction in dimension means algorithms must make choices and disregard certain information. In this thesis, we improve upon existing dimension reduction methods through the use of hyper-dimensional spanning trees to preserve complex clustering relationships in the original data. We demonstrate the approach with an interactive program for our dimension reduction algorithm, which allows the user to fine-tune the outcome.
Degree
thesis:*- Name thesis:degree_name
- MS in Computer Science (M.S.C.S.)
- Level thesis:degree_level
- Thesis
- Discipline thesis:degree_discipline
- Computer Science and Engineering
- Grantor
- University of South Florida
- Year dc:date
- 2021
Author and committee
dc:creator, dc:contributor.*- Author dc:creator
-
- Davis, Curtis Thomas
- Contributors dc:contributor
-
- Paul Rosen, Ph.D.
- Shaun Canavan, Ph.D.
- Tempestt Neal, Ph.D.
Subjects
dc:subject × 6Identifiers
dc:identifier.*- Repository record dc:identifier
- https://digitalcommons.usf.edu/etd/9096