Problem solved.
Thomas Neal has helpfully pointed out that the OpenBUGS code now "checks
the zero-sum constraint on initial values and also checks the wts and
other inputs more carefully for consistency with the specified input
requirements."
I wasn't generating my initial values carefully enough. Older versions
of OpenBUGS and WinBUGS ran happily even though the initial values I had
supplied were nonsensical.
best wishes
Mark
-------- Original Message --------
Subject: Re: [BUGS] Spatial car.normal initial value errorþ
Date: Tue, 21 Sep 2010 09:44:26 +0100
From: Mark Strong <[log in to unmask]>
Reply-To: Mark Strong <[log in to unmask]>
To: [log in to unmask]
Dear all
I have discovered the same "error for node f[1] of type
SpatialCARNormal.Node node has invalid value error pos 70737 (error on
line 1)" problem using OpenBUGS 3.1.1 on both Linux and Windows platforms.
OpenBUGS 3.0.3 (and therefore the current version of BRugs I guess) does
not generate the error, and neither does WinBUGS.
This seems to be a particular problem with the newer (post 3.0.3)
versions of OpenBUGS.
I would be careful using OpenBUGS for spatial models - apparently there
were known issues in previous versions also. It would be good to know
when these are resolved.
many thanks to the developers.
Mark
On 21/09/2010 05:05, Hsieh Angela wrote:
>
> Hi,
>
> I’m running my spatial model by using OpenBUGS on Linux. I got the
> following error when I tried to initialize several chains: “error for
> node f[1] of type SpatialCARNormal.Node node has invalid value error
> pos 70737 (error on line 1)”
>
> Here, f is a vector that follows the CAR normal model with the form:
> f[1:N] ~ car.normal(adj[], wts[], num[], tau)
>
> But if I set all the initial values for f to be 0, I would not get any
> error message. However, if I change any initial values for f not equal
> to 0, then I’ll get above error message.
>
> Actually, I’ve also tried to run the same model, dataset, and initial
> values on WinBUGS and BRugs in R on Windows. And I didn’t get any
> error message by using either of them on Windows. But I don’t know why
> I would get an error message associated with the parameters with CAR
> normal model when I try to run it on OpenBUGS for Linux,
>
> Does anybody have any idea what might be causing that error? Thank you
> very much!
>
> Angela
>
> -------------------------------------------------------------------
> 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]<mailto:[log in to unmask]> To mail the
> BUGS list, mail to [log in to unmask]<mailto:[log in to unmask]>
> Before mailing, please check the archive at
> www.jiscmail.ac.uk/lists/bugs.html
> <http://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]<mailto:[log in to unmask]> If this
> fails, mail [log in to unmask]
> <mailto:[log in to unmask]>, NOT the whole list
--
Mark Strong
MRC Fellow
Section of Public Health
School of Health and Related Research
University of Sheffield
Regent Court
30 Regent Street
Sheffield S1 4DA
Tel: 0114 222 0812
Fax: 0114 272 4095
[log in to unmask]
http://www.shef.ac.uk/scharr/sections/ph/staff/profiles/mark.html
-------------------------------------------------------------------
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
-------------------------------------------------------------------
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
|