![Electronic Theses and Dissertations](../../assets/md5images/8bf39a2f649bfcc97f3e39e2323994d5.png)
Date of Award
1-1-2023
Document Type
Dissertation
Degree Name
Ph.D. in Engineering Science
First Advisor
Yixin Chen
Second Advisor
Thai Le
Third Advisor
Charlie Walter
Relational Format
dissertation/thesis
Abstract
Identifying and creating patterns enables humans to fulfill their fundamental desire to make sense of things. At the center of this endeavor, seriation is a process of arranging objects to reveal regularities in their characteristics. We present a method to arrange objects cyclically to achieve orderliness based on a novel concept of perturbation entropy, which models the randomness of a sequence under local perturbations on the ordering. We derive a sharp upper bound on the perturbation entropy and develop an algorithm to search for an ordering that minimizes the upper bound. Connections with the traveling salesman problem and the quadratic assignment problem are discussed and explored. Extensive empirical studies demonstrate the performance of the proposed approach in a variety of applications and the robustness of the algorithm against complex topologies and noise induced by the sparsity of data.
Recommended Citation
Rego, Carla, "ROCS: Robust Circular Seriation by Perturbation Entropy Minimization" (2023). Electronic Theses and Dissertations. 2767.
https://egrove.olemiss.edu/etd/2767