On October 7 I raised the possibility of organising a BUGS workshop. To
date I have received about 25 favourable replies, some individuals, some
representing groups of statisticians and others from the US whom I
assume will not attend a meeting in the UK. I believe there is enough
interest to proceed with organising such a meeting, especially as there
is likely to be others outside the BUGS mailing list who would be also
interested.
I have provisionally booked Errol St. (RSS HQ in London) for Thursday
June 24.
Both David Spiegelhalter and Nicky Best have agreed to take part.
My suggested working title for this meeting is "Practical Bayes using
BUGS".
Personally I would wish this workshop to be predominantly about
practical application issues. Examples might include:
How to use DOODLEBUGS (advantages?)
Assessing convergence
Model validation
Specifying "weak" priors in "non-standard" but reasonably common
situations
Understanding error messages, common mistakes, FAQ's, "tricks &
tips"
Practical issues with non log-concave distributions (?)
Effect & implications of different parameterisations - good & bad
practice
Some specific issues that were raised: multinomial probit models,
BMA
Some of these issues could be discussed via poster presentations.
My current thoughts on the format of the meeting are to base the day
around 4,say, breakout small discussion groups - each on a specific
topic. Each group then presents back to the whole in the afternoon;
these "presentations" could take many forms: problems posed and
tentative solutions offered for discussion, good/bad egs shared, "cries
for help", ... Concurrent with this could be the poster presentations.
It would be nice to finish the day with a brief talk on future BUGS
developments.
We could also use the day to obtain feedback on the additional BUGS
features that users would like, and current features not liked.
Alternatively this could be the main focus of the day.
The meeting will be advertised in RSS News and allstat March/April next
year.
If you have any comments on my proposals, or suggestions for wider
advertising I would very much welcome them. It would be particularly
useful to receive "votes", from likely attendees, on which topics are of
interest and whether the proposed format is appropriate (alternative
suggestions are encouraged). Also, if you wish to offer a poster
presentation please forward me a synopsis, preferably by end of February
'99.
To avoid general annoyance, please reply direct to me and not the
mailing list. I will update the list on progress early in the new year.
Regards
Phil
LEGAL NOTICE
Unless expressly stated otherwise, this message is confidential and
may be privileged. It is intended for the addressee(s) only. Access
to
this e-mail by anyone else is unauthorised. If you are not an
addressee, any disclosure or copying of the contents of this e-mail
or
any action taken (or not taken) in reliance on it is unauthorised
and
may be unlawful. If you are not an addressee, please inform the
sender
immediately.
Pfizer Limited is registered in England under No. 526209 with its
registered office at Ramsgate Road, Sandwich, Kent CT13 9NJ
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|