Date of Award

2023

Degree Type

Thesis

Degree Name

Master of Science in Statistics

Department

Computer Science and Statistics

First Advisor

Jing Wu

Abstract

This project aimed to investigate the differences in terms of network structural characteristics between full and sampled networks and to develop a new weighted generalized estimating equations model (WGEE) in order to mitigate the sampling bias and obtain less biased sample estimates for model coefficients and standard errors. The applications of such an approach are of interest in the field of network analysis and public health studies. The Transmission Reduction Intervention Project (TRIP) is an HIV study using risk networks and is a key motivator for this work. Two other motivating examples were constructed where it was found that sampled networks often do not share similar values for some key network characteristics. To attempt to reduce bias induced by these sampled networks, multiple weight functions were applied to each potential edge using four different approaches. Then a weighted generalized estimating equations (WGEE) model was fit on a simulated full network and a sample from that network using a computationally efficient one-step WGEE method. The results of this simulation show that bias is present in the unweighted case when compared to what is observed in the full network. Weighting approaches show a promising direction for future work on addressing bias in WGEE models in networks.

Creative Commons License

Creative Commons Attribution 4.0 License
This work is licensed under a Creative Commons Attribution 4.0 License.

Share

COinS
 
 

To view the content in your browser, please download Adobe Reader or, alternately,
you may Download the file to your hard drive.

NOTE: The latest versions of Adobe Reader do not support viewing PDF files within Firefox on Mac OS and if you are using a modern (Intel) Mac, there is no official plugin for viewing PDF files within the browser window.