Re: Gibbs sampling and annealing (was: Help! Book recommendations?)

Article View: sci.math.stat

Article #648

#648

From: Charles Geyer
Date: Tue, 06 Oct 1992 22:52

27 lines
1640 bytes

In article <1992Oct5.225003.11916@news2.cis.umn.edu> I wrote:

  [Annealing] doesn't help much.  All it attempts to do is produce one
  realization that is "almost" from the stationary distribution to
  use as the start of the run.

  But if the chain mixes very slowly, the average will be near the starting
  point, but not near the expectation one wants to calculate.

  On the other hand, if the chain mixes rapidly, the "annealing" isn't
  needed.

In article <92Oct6.115856edt.512@neuron.ai.toronto.edu> radford@cs.toronto.edu
(Radford Neal) writes:

> Not necessarily. It could be that the chain mixes rapidly once it has
> reached the right general region, but that without annealing it takes
> a long time to reach that region from whereever it was started.

Possible, in which case I'll concede that annealing to find starting
points is a good idea.

-- 
Charles Geyer
School of Statistics
University of Minnesota
charlie@umnstat.stat.umn.edu

Technical Details

Message-ID: <1992Oct6.225231.25621@news2.cis.umn.edu>

Path: arc1-ca.pugleaf.net!sparky!uunet!think.com!sdd.hp.com!spool.mu.edu!umn.edu!charlie

References:

<92Oct5.123636edt.472@neuron.ai.toronto.edu> <1992Oct5.225003.11916@news2.cis.umn.edu> <92Oct6.115856edt.512@neuron.ai.toronto.edu>

Show Full Headers

Complete Article Headers

Newsgroups: sci.math.stat,sci.math.num-analysis
Path: arc1-ca.pugleaf.net!sparky!uunet!think.com!sdd.hp.com!spool.mu.edu!umn.edu!charlie
From: charlie@umnstat.stat.umn.edu (Charles Geyer)
Subject: Re: Gibbs sampling and annealing (was: Help!  Book recommendations?)
Message-ID: <1992Oct6.225231.25621@news2.cis.umn.edu>
Sender: news@news2.cis.umn.edu (Usenet News Administration)
Nntp-Posting-Host: isles.stat.umn.edu
Organization: School of Statistics, University of Minnesota
References: <92Oct5.123636edt.472@neuron.ai.toronto.edu> <1992Oct5.225003.11916@news2.cis.umn.edu> <92Oct6.115856edt.512@neuron.ai.toronto.edu>
Date: Tue, 6 Oct 1992 22:52:31 GMT
Lines: 27

Reply to Parent Message