Abstract
dc:description.abstractWith the quick progression of technology and the increasing need to process large data, there has been an increased interest in data-dependent and data-independent dimension reduction techniques such as principle component analysis (PCA) and Johnson\-Lindenstrauss (JL) transformations, respectively. In 1984, Johnson and Lindenstrauss proved that any finite set of data in a high-dimensional space can be projected into a low-dimensional space while preserving the pairwise Euclidean distance within any desired accuracy, provided the projected dimension is sufficiently large; however, if the desired projected dimension is too small, Woodruff and Jayram, and Kane, Nelson, and Meka in 2011 separately proved such a projection does not exist. In this thesis, we answer an open problem by providing a precise threshold for the projected dimension, above which, there exists a projection approximately preserving the Euclidean distance, but below which, there does not exist such a projection. We, also, give a brief survey of JL constructions, covering the initial constructions and those based on fast-Fourier transforms and codes, and discuss applications in which JL transformations have been implemented.
Degree
thesis:*- Name thesis:degree_name
- Doctor of Philosophy (PhD)
- Level thesis:degree_level
- Dissertation
- Discipline thesis:degree_discipline
- Mathematical Sciences
Author and committee
dc:creator, dc:contributor.*- Author dc:creator
-
- Knoll, Fiona
- Contributors dc:contributor
-
- Dr. Shuhong Gao, Committee Chair
- Dr. Gretchen Matthews
- Dr. Felice Manganiello
- Dr. June Luo
Subjects
dc:subject × 3Identifiers
dc:identifier.*- Repository record dc:identifier
- https://open.clemson.edu/all_dissertations/1977