New Summit website coming in May 2021!

                   Check the SFU library website for updates.

Bayesian logistic regression with the local bouncy particle sampler for COVID-19

Author: 
Date created: 
2020-08-24
Identifier: 
etd21018
Keywords: 
COVID-19
Logistic regression model
Markov chain Monte Carlo
Bouncy particle sampler
Local bouncy particle sampler.
Abstract: 

A novel coronavirus, called SARS-CoV-2, has caused the outbreak of the pandemic of COVID-19. The global economy, people’s health and life have been facing a tremendous threat in COVID-19. This project is to determine some important factors in COVID-19 severity based on 137 Tianjin patients who have been exposed to COVID-19 since January 5, 2020. We fit a logistic regression model and estimate the parameters using standard Markov chain Monte Carlo (MCMC) methods. Due to the weaknesses and limitations of the standard MCMC methods, we then perform model estimation in one special example of a Piecewise Deterministic Markov Process, named the Bouncy Particle Sampler (BPS). This method is also known as a rejection-free and irreversible MCMC, and can draw samples from our target distribution efficiently. One type of the BPS algorithm, the Local Bouncy Particle Sampler (LBPS), has advantages in computational efficiency. We apply the standard MCMC method and the LBPS to our dataset. We conclude that age and Wuhan-related exposures (i.e. people who have lived or traveled from Wuhan) are two important factors in a COVID-19 severity test.

Document type: 
Graduating extended essay / Research project
Rights: 
This thesis may be printed or downloaded for non-commercial research and scholarly purposes. Copyright remains with the author.
File(s): 
Supervisor(s): 
Liangliang Wang
Department: 
Science: Department of Statistics and Actuarial Science
Thesis type: 
(Project) M.Sc.
Statistics: