Abstract
dc:description.abstractThe dissertation aims to address the problem of tail index estimation for stationary time series with regularly varying marginal distributions, a topic extensively studied for independent data and certain types of dependent data. We consider to obtain the asymptotic properties of the tail index estimator under the framework of a recently proposed concept of serial tail dependence known as tail adversarial stability. This work focus on the Hill’s estimator, a widely used tail index estimator, within a broad class of tail-dependent time series. To tackle this problem, we employ the technique of adversarial innovation coupling combined with a martingale approximation scheme to develop the consistency and central limit theorem for Hill’s estimator for a general class of tail dependent time series. Leveraging the asymptotic properties, we propose an application involving the clustering of a large number of regularly varying time series based on their tail indices through a robust mixture algorithm, which is valuable for identifying and analyzing subgroups of time series that share similar extremal behavior. We also conduct Monte Carlo simulations for moving-maximum processes to examine the finite sample performance and present real data analyses.
Degree
thesis:*- Name thesis:degree_name
- Doctor of Philosophy
- Level thesis:degree_level
- doctoral
- Discipline thesis:degree_discipline
- Mathematics & Statistics
- Grantor
- Boston University
- Year dc:date.issued
- 2025
Author and committee
dc:creator, dc:contributor.*- Author dc:creator
-
- Cao, Hanyue
- Advisor dc:contributor.advisor
-
- Zhang, Ting
Subjects
dc:subject × 4Rights
dc:rights- Statement dc:rights
-
- Attribution 4.0 International
- Licence dc:rights.uri
- Language dc:language.iso
- en_US
Identifiers
dc:identifier.*- Handle dc:identifier.uri
- https://hdl.handle.net/2144/53330