Georgia Institute of Technology
Statistical learning and change detection for dynamic networks
Abstract
dc:description.abstractThe field of graph analysis has recently garnered a significant amount of attention due to its importance in maximizing the utility of data. While most prior research in graph learning has primarily focused on static graphs with fixed nodes and edges, there are numerous real-world applications where the graph structure evolves dynamically, such as social networks, sales data, or transportation systems. In this thesis, we concentrate on a specific type of graph, called the Dynamic Graph, which aims to capture such graph dynamics. Our study commences with an exploration of discrete-time dynamic graphs and identifies structural changes based on a statistical model utilizing a spectral CUSUM method. Furthermore, we examine continuous- time dynamic graphs to model solar power in a practical solar system and forecast anomalous solar activities in real-time using a marked temporal point process model. In the end, we investigate the financial impact of serious environmental violations committed by suppliers on purchasing firms by analyzing supply chain networks that link the suppliers with the buying firms.
Degree
thesis:*- Level thesis:degree_level
- Doctoral
- Department dc:contributor.department
- Industrial and Systems Engineering
- Grantor dc:publisher
- Georgia Institute of Technology
- Year dc:date.issued
- 2023
Author and committee
dc:creator, dc:contributor.*- Author dc:creator
-
- Zhang, Minghe
- Advisor dc:contributor.advisor
-
- Xie, Yao
- Committee members dc:contributor.committeemember
-
- Van Hentenryck, Pascal
- Paynabar, Kamran
- Kalkanci, Basak
- Qiu, Feng
Subjects
dc:subject × 4Rights
- Language dc:language.iso
- en_US
Identifiers
dc:identifier.*- Handle dc:identifier.uri
- https://hdl.handle.net/1853/71973