A sequential Monte Carlo approach to joint longitudinal and time-to-event modeling
Abstract/Contents
- Abstract
- The statistical modeling framework known as joint longitudinal/time-to-event modeling (joint modeling) has enabled researchers to combine the benefits of two important modeling paradigms in a unified fashion, as it aims to capture the interrelatedness of the respective models while delivering improvements in efficiency and inferential validity. Much of this work has evolved in the biostatistical setting, where researchers wish to account for the interrelationship between dynamically-evolving biological processes and survival time distributions. The vast majority of estimation techniques for joint modeling have been developed in the maximum-likelihood framework. While Bayesian analyses do exist, their practicality is hindered by the typical computational drawbacks associated with classical Markov Chain Monte Carlo estimation techniques, which do not generalize to longitudinal settings in which one may desire to perform online analyses. In this thesis, after a comprehensive review of joint modeling, we describe a new approach to Bayesian estimation of general state space models which provides a solution to the joint state and parameter estimation in an online fashion. As a means to this end, we describe the development of a computationally efficient alternative to traditional Bayesian inference by Markov Chain Monte Carlo (MCMC) methods, and explain how it can be incorporated into a novel adaptive sequential Monte Carlo framework to allow for fast and accurate estimation of model states and parameters. We then describe how a general joint longitudinal/time-to-event model may be constructed in a state space context, and carry out detailed simulation studies to illustrate the performance of the algorithm and utility of the model. We conclude that approaching joint longitudinal/time-to-event models through the prism of state space modeling offers a new and constructive perspective on an important modern statistical framework.
Description
| Type of resource | text |
|---|---|
| Form | electronic; electronic resource; remote |
| Extent | 1 online resource. |
| Publication date | 2015 |
| Issuance | monographic |
| Language | English |
Creators/Contributors
| Associated with | Gewitz, Andrew David |
|---|---|
| Associated with | Stanford University, Institute for Computational and Mathematical Engineering. |
| Primary advisor | Lai, T. L |
| Thesis advisor | Lai, T. L |
| Thesis advisor | Lavori, Philip W, 1949- |
| Thesis advisor | Shih, Mei-Chiung |
| Advisor | Lavori, Philip W, 1949- |
| Advisor | Shih, Mei-Chiung |
Subjects
| Genre | Theses |
|---|
Bibliographic information
| Statement of responsibility | Andrew David Gewitz. |
|---|---|
| Note | Submitted to the Institute for Computational and Mathematical Engineering. |
| Thesis | Thesis (Ph.D.)--Stanford University, 2015. |
| Location | electronic resource |
Access conditions
- Copyright
- © 2015 by Andrew David Gewitz
Also listed in
Loading usage metrics...