RE: Output of each chain in Bayses and ISAMPLE >1

From: Robert Bauer Date: March 04, 2011 technical Source: mail-archive.com
MCMC automatically sets ISAMPLE=1 for its run. ISAMPLE>1 is only for SAEM. To perform multiple chains, you must run separate control stream problems with different initial SEED settings.
Quoted reply history
-----Original Message----- From: [email protected] [mailto:[email protected]] On Behalf Of Dieter Menne Sent: Friday, March 04, 2011 3:44 AM To: [email protected] Subject: [NMusers] Output of each chain in Bayses and ISAMPLE >1 Dear nmusers, To compute the DIC from using Plummer's definition, I need the output of multiple chains in BAYES. With ISAMPLE=2, I did not get an mcmc trace that was different from the default of ISAMPLE=1. Do I have to add some other statement for separate printing? The only mentioning of the word "chain" in the complete documentation I found was in context of ISAMPLE. Plummer, M. (2002), Discussion of the paper by Spiegelhalter et al. Journal of the Royal Statistical Society Series B 64, 620. Dieter Menne
Mar 04, 2011 Dieter Menne Output of each chain in Bayses and ISAMPLE >1
Mar 04, 2011 Robert Bauer RE: Output of each chain in Bayses and ISAMPLE >1