Multi-scale persistent homology

UNCG Author/Contributor (non-UNCG co-authors, if there are any, appear on document)
Austin Lawson (Creator)
Institution
The University of North Carolina at Greensboro (UNCG )
Web Site: http://library.uncg.edu/
Advisor
Gregory Bell

Abstract: Data has shape and that shape is important. This is the anthem of Topological Data Analysis (TDA) as often stated by Gunnar Carlsson. In this paper we take a common method of persistence involving the growing of balls of the same size, and generalizing to balls of different sizes in order to better understand the outlier and coverage problems. We begin with a summary of classical persistence theory and stability. We then move on to generalizing the Rips and \v{C}ech complexes as well as generalizing the Rips lemma. We transition into 3 notions of stability in terms of bottleneck distance. For the outlier problem, we show that it is possible to interpolate between persistence on a set with no noise and a set with noise. For the coverage problem, we present an algorithm which provides a cheap way of covering a compact domain.

Additional Information

Publication
Thesis
Language: English
Date: 2016
Keywords
Complex, Coverage, Homology, Persistent, Stability
Subjects
Homology theory
Computational complexity
Topology

Email this document to