Multilevel Model for High Dimensional Longitudinal Paired Data

Authors

Manuel Roosevelt LampteyFollow

Document Type

Thesis

Degree

Master of Science (MS)

Major/Program

Statistics

First Advisor's Name

Wensong Wu

First Advisor's Committee Title

Committee Chair

Second Advisor's Name

Ian L. Dryden

Second Advisor's Committee Title

Committee Member

Third Advisor's Name

Florence George

Third Advisor's Committee Title

Committee Member

Keywords

biostatistics, clinical trials, longitudinal data analysis and time series, statistics and probability

Date of Defense

6-16-2023

Abstract

Longitudinal data, which consists of repeated measurements on the same subjects over time, is very common in many scientific settings. The main goal is to describe and understand how the response variable changes over time. In some settings, high dimensional or even unstructured data are collected in pairs of treatments on the same subject at each time point. This thesis developed a method to analyze such high dimensional longitudinal paired data by a two-step procedure: First reduce the dimension of the data, and then fit multilevel models to estimate the treatment and the time effect. This method was applied to a longitudinal study where imaging data was collected on breast cancer patients undergoing radiation therapy. Various dimension reduction methods and multilevel models varied in their results and performance. The findings indicated a statistically significant alteration in tissue oxygenation levels over time and a distinction between the radiated and control breasts.

Identifier

FIDC011219

ORCID

https://orcid.org/0009-0006-1552-5362

Recommended Citation

Lamptey, Manuel Roosevelt, "Multilevel Model for High Dimensional Longitudinal Paired Data" (2023). FIU Electronic Theses and Dissertations. 5346.
https://digitalcommons.fiu.edu/etd/5346