Date of Award
2024
Degree Type
Thesis
Degree Name
Master of Science in Computer Science
Department
Computer Science and Statistics
First Advisor
Noah Daniels
Abstract
Visualizing high dimensional data can be a challenging task due to the difficulty people face in comprehending information beyond three dimensions. Further research and development of tools in this area could prove valuable for creating efficient, intuitive, and accurate visualizations. It could also provide insight into the manifold hypothesis, which suggests that high dimensional data can exist in low dimensional space.
This thesis proposes the utilization of clustered manifold mapping as a novel visualization technique that summarizes a dataset into a hierarchal tree of clusters by partitioning the data based on a user-specified distance metric. A subset of clusters can be carefully selected from the tree to create a 3D graph using the Unity game engine, which enables the user to interact with the and explore various features of the data.
The graphs produced with this approach will be quantitatively and qualitatively compared with existing methods such as UMAP, which demonstrates its contributions to the field of visualizing high dimensional data. Furthermore, visualizing the tree of clusters in addition to the graph provides a greater understanding into the field of clustered manifold mapping.
Recommended Citation
Perrone, David, "VISUALIZATION OF HIGH DIMENSIONAL DATA IN LOW DIMENSIONAL SPACE VIA CLUSTERED MANIFOLD MAPPING" (2024). Open Access Master's Theses. Paper 2500.
https://digitalcommons.uri.edu/theses/2500