Next: Some econometric duration
Up: Latent Variable Models for
Previous: Overview of the
- Casella and George(1992); Smith and Roberts(1993)
- Parameters of interest are
.
- Data to hand are
.
- Latent data are
.
- Basis of inference is the likelihood
.
- Or, what is nearly the same, the posterior,
.
- Make inferences about
by sampling 
.
- Or, equivalently, sampling
- Why equivalently? Think about it.
- How to sample
?
- Sample, in sequence, the component conditionals.
- These are
-
-
- 1. is simple --- see the definition of latent variables, slide 1!!
- 2. is usually simple.
- One pass through the component conditionals is one step.
Do many steps. Later
realizations come from
, as required.
Tony Lancaster
tl@pstc3.pstc.brown.edu
Last modified: 12 September 1995