Dear Alstat,
As you may know, some days ago, I posted an inquiry to the list to get help
on implementation of Winbugs in dynamic Bayesian models. However, I got few
comments, some of them are not helpful. If you are interested, look at the
following replies!
With the best wishes,
ALI
=========================================================================
-----Original Message-----
From: Andrew Millard [mailto:[log in to unmask]]
Sent: Wednesday, December 10, 2003 5:40 PM
To: Ali Bromideh
Subject: Re: [BUGS] Correction!!
Dear Ali
I find that dynamic linear models do not work well in WinBUGS, as they
take a long time to converge. I have an example for a simple random walk
model of the radiocarbon calibration curve on my webpage:
http://www.dur.ac.uk/a.r.millard/BUGS4Arch.html under Christen JA &
Nicholls G (2000).
If your model is not too complex then calculation in R/S+ without MCMC is
easier - see for example the R code of Jonty Rougier for time-series
dynamic linear models
http://www.maths.dur.ac.uk/stats/people/jcr/softR.html (scroll down to
tsdlm)
Andrew
=========================================================================
Dr. Andrew Millard [log in to unmask]
Department of Archaeology, University of Durham, Tel: +44 191 334 1147
South Road, Durham. DH1 3LE. United Kingdom. Fax: +44 191 334 1101
http://www.dur.ac.uk/a.r.millard/
=========================================================================
-----Original Message-----
From: [log in to unmask] [mailto:[log in to unmask]]
Sent: Tuesday, December 09, 2003 9:24 PM
To: Ali Bromideh
Subject: Re: [BUGS] Inquiry.
I am not quite sure about what counts as dynamic. I have an example of
adaptative process control (one responses, two control factors). The
algorithm much like Kalman's filter, with ever-changing coefficients. Its
claim to novelty, such as it is, comes from the fact that the control
uncertainty inflates as the time since last monitor point increases. From
the one-dimensional family of control prescriptions, one prescription has
uniquely smallest prediction variance, and is selected as the feedback
control adjustment.
Based on the IKCO, this looks like a relevant application to you. In the
context of pharmaceutical investigations, similar ideas can be applied to
pharmacokinetics (without the feedback).
______________________________________________________________________
-----Original Message-----
From: Kenneth Rice [mailto:[log in to unmask]]
Sent: Tuesday, December 09, 2003 5:14 PM
To: Ali Bromideh
Cc: Paivi Putkonen
Subject: Re: [BUGS] Inquiry.
Ali,
Your question seems popular - see below for someone else interested in
dynamic
models. Please make sure you post a summary of all the (useful) replies you
get;
just paste them together and mail to [log in to unmask]
Many thanks
Ken Rice
Bugs list manager
_____________________________________________________
Dr Kenneth Rice
MRC Biostatistics Unit
Inst. of Public Health
Forvie Site, Robinson Way
Cambridge, CB2 2SR
Tel: (+44) 01223-330-369
www.statslab.cam.ac.uk/~krice
------------- Begin Forwarded Message -------------
X-RAL-MFrom: <[log in to unmask]>
X-RAL-Connect: <[log in to unmask] [130.233.228.94]>
Date: Tue, 9 Dec 2003 13:51:34 +0200 (EET)
From: Paivi Putkonen <[log in to unmask]>
cc: [log in to unmask]
Subject: Re: [BUGS] Inquiry.
MIME-Version: 1.0
X-RAVMilter-Version: 8.4.3(snapshot 20030212) (smtp-4.hut.fi)
X-Scanned-By: MIMEDefang 2.38
Content-Transfer-Encoding: quoted-printable
X-MIME-Autoconverted: from 8bit to quoted-printable by fili.jiscmail.ac.uk
id
hB9BpbfX001934
Hi,
I?m also interested in anyones experiences of dynamic and
non-dynamic Markov models with winbugs, while looking forward to choose a
system to run several models with a relatively small patient data.
With Respect
Researcher
Paivi Putkonen
________________________________________________________________________
Paivi Putkonen M.Sc (Health Sciences) Research Scientist
Helsinki University of Technology Tel: +358-9-451 3180
Laboratory of Biomedical Engineering Fax: +358-9-451 3182
P.O.Box 2200 FIN-02015 HUT Finland
email: [log in to unmask]
______________________________________________________________________
-------------------------------------------------------------------
This list is for discussion of modelling issues and the BUGS software.
For help with crashes and error messages, first mail [log in to unmask]
To mail the BUGS list, mail to [log in to unmask]
Before mailing, please check the archive at www.jiscmail.ac.uk/lists/bugs.html
Please do not mail attachments to the list.
To leave the BUGS list, send LEAVE BUGS to [log in to unmask]
If this fails, mail [log in to unmask], NOT the whole list
|