Estimation Approaches of Sufficient Reduction under Kronecker Product Structure

We propose methods to estimate linear combinations of several predictors that aremeasured repeatedly over time, that contain sufficient information for the regressionof the predictors on the outcome. We assume that the first moments of the predictorscan be decomposed into a time and a marker component via a Kronecker productstructure that accommodates the longitudinal nature of the predictors. We thenextend least squares and model-based sufficient dimension reduction techniques toaccommodate this setting. We derive some analytic properties, and we compare theperformance of the various approaches under different assumptions on the structureof the predictors in simulations.

Author

• Wang, Wei

Language

• en

Keyword

• Sufficient dimension reduction
• Kronecker product
• Longitudinal data
• Principal fitted components

Date created

• 2018

Type of Work

• Master's Thesis

License

• All rights reserved

Rights statement

• In Copyright

GW Unit

• Statistics

Degree

• M.S.

Advisor

• Pfeiffer, Ruth
• Bura, Efstathia

Persistent URL

•  https://scholarspace.library.gwu.edu/etd/9s161646v