KIT | KIT-Bibliothek | Impressum | Datenschutz
Open Access Logo
DOI: 10.5445/IR/1000068981

Change-Point Procedures for Multivariate Dependent Data

Weber, Silke Martina

Abstract (englisch):
The change-point procedure searchs for structural changes in data collected over time, for example a change in the mean or variance. We consider multivariate data which can be dependent.
Precisely, block bootstrap methods for classical tests for changes in the mean are investigated as well as new test statistics introduced, where -- unlike in the classical setting -- the location of the changes can be at different locations in different components.
In the sequential setting, some unifying theory based on estimating functions is developped extending the currently available methodology in several directions.

Zugehörige Institution(en) am KIT Institut für Stochastik (STOCH)
Publikationstyp Hochschulschrift
Jahr 2017
Sprache Englisch
Identifikator URN: urn:nbn:de:swb:90-689812
KITopen-ID: 1000068981
Verlag Karlsruhe
Umfang 213 S.
Abschlussart Dissertation
Fakultät Fakultät für Mathematik (MATH)
Institut Institut für Stochastik (STOCH)
Prüfungsdatum 08.02.2017
Referent/Betreuer Prof. C. Kirch
Schlagworte time series, change-point, bootstrap
KIT – Die Forschungsuniversität in der Helmholtz-Gemeinschaft KITopen Landing Page