-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 "Devon McCormick" <[EMAIL PROTECTED]> writes:
> I am attempting a Bayesian solution to the problem. As is often the > case with this method, the hard part is coming up with a good prior > probability. MCMC is another possibility I may look at. I looked at > WinBUGS a long time ago - whether or not I re-visit it may depend on > how other approaches go. Devon, FWIW, I'm using MCSim to do system dynamics modeling these days, and it includes MCMC capability. The pertinence to J is that I'm using J as the tool to analyze MCSim output. See http://www.facilitatedsystems.com/weblog/2006/11/system-dynamics-with-mcsim.html for links to a presentation I did on the matter and a quick reference card that gives a short (and probably not so excellent) J script to load and process some MCSim data. One of my next steps in that regard will be to write scripts to stuff other forms of MCSim output into J arrays, probably of the same structure as those from the script on the quick reference card to make it a bit less confusing to analyze. I'm in the process of learning enough about MCMC to add that to my repertoire (pointers to more useful information are welcome!). In the process, I found but have yet to install First Bayes (http://www.tonyohagan.co.uk/1b/), a learning application written in Dyalog APL/W. Perhaps we should encourage its author, Tony O'Hagan, to provide First Bayes as a J lab! Follow-ups may more properly belong in J Chat. Bill - -- Bill Harris http://facilitatedsystems.com/weblog/ Facilitated Systems Everett, WA 98208 USA http://facilitatedsystems.com/ phone: +1 425 337-5541 -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.3 (MingW32) Comment: For more information, see http://www.gnupg.org iD8DBQFFrROT3J3HaQTDvd8RAn7tAJ9vtOm+AMN8edMJA3SRsi4DQVjT5ACferP5 kDed/0kjdIh6Swg7xBdDhtg= =XtF9 -----END PGP SIGNATURE----- ---------------------------------------------------------------------- For information about J forums see http://www.jsoftware.com/forums.htm
