Introducing Markov chain Monte Carlo
- 1 December 1995
- book chapter
- Published by Taylor & Francis
- p. 19-38
- https://doi.org/10.1201/b14835-6
Abstract
Most applications of MCMC to date, including the majority of those described in the following chapters, are oriented towards Bayesian inference. From a Bayesian perspective, there is no fundamental distinction between The integrations in this expression have until recently been the source of most of the practical difficulties in Bayesian inference, especially in high dimensions. In most applications, analytic evaluation of E[J(0)ID] is impossible. Alternative approaches include numerical evaluation, which is difficult and inaccurate in greater than about 20 dimensions; analytic approximation such as the Laplace approximation (Kass et al., 1988), which is sometimes appropriate; and Monte Carlo integration, including MCMC.Keywords
This publication has 1 reference indexed in Scilit:
- Spatial Interaction and the Statistical Analysis of Lattice SystemsJournal of the Royal Statistical Society Series B: Statistical Methodology, 1974