## Taking Bigger Metropolis Steps by Dragging Fast
Variables

**Radford M. Neal,
Dept. of Statistics and Dept. of Computer Science, University of Toronto**

I show how Markov chain sampling with
the Metropolis-Hasting algorithm can be modified so as to take bigger
steps when the distribution being sampled from has the characteristic
that its density can be quickly recomputed for a new point if this
point differs from a previous point only with respect to a subset of
``fast'' variables. I show empirically that when using this method,
the efficiency of sampling for the remaining ``slow'' variables can
approach what would be possible using Metropolis updates based on the
marginal distribution for the slow variables.

Technical Report No. 0411, Dept. of Statistics, University of Toronto
(October 2004, typos corrected February 2005), 9 pages:
postscript,
pdf.

Also available
from arXiv.org.

You can also get the program used for the
tests in this paper.