ST955: Dissertation, MSc Statistics, University of Warwick
Markov melding, proposed in the paper Joining and splitting models with Markov melding by Robert Goudie and colleagues, is a Bayesian computational method for evidence synthesis.
This masters dissertation is an outline of the method, together with an implementation of the algorithm in R
applied to a simulated data example.
- Code to reproduce all figures, tables etc. is contained within the R Markdown files for each chapter e.g.
meld/thesis/02-modelling.Rmd
for Chapter 2. The final thesis is uploaded here. - For Example 3.2 the Bayesian random-effects meta-analysis is fit using
meld/code/smith_mwg.R
, Bayesian fixed-effects meta-analysis usingmeld/code/fixed_mwg.R
and Markov melded model usingmeld/code/meld_mwg.R
.