UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Markov chain Monte Carlo algorithm comparisons Wen, Sijin

Abstract

Various Markov chain Monte Carlo algorithms are available for sampling from a posterior distribution. The random walk Metropolis algorithm is a simple scheme which is frequently used in Bayesian statistical problem. The guided walk algorithm attempts to suppress the random walk behavior in the random walk Metropolis algorithm. Other algorithms, such as the Langevin algorithm and the hybrid algorithm use more information about the posterior distribution than the random walk Metropolis algorithm and the guided walk algorithm. In this thesis, The performance of each of those four algorithms has been examined, based on simulation studies using multivariate normal target distributions. Then we compare the algorithms in terms of efficiency and convergence time. Moreover, these four algorithms are compared using a posterior distribution for parameters given observed data in an application.

Item Media

Item Citations and Data

Rights

For non-commercial purposes only, such as research, private study and education. Additional conditions apply, see Terms of Use https://open.library.ubc.ca/terms_of_use.

Usage Statistics