Re: [UAI] Required sample size for parameter and structure learning

Subject:

Comments:
: : Here is a good link to this topic: : : http://www.cs.ust.hk/~samee/bayesian/bayes.html : Thanks, Haipeng. : The Friedman and Yakhini paper (http://www.cs.ust.hk/~samee/bayesian/papers/k.ps)discusses a very powerful MDL-based computational learning theoretic (COLT) approach. : I remember the Greiner, Grove, and Schuurmans paper (http://www.cs.ust.hk/~samee/bayesian/papers/ggs-bnpw-uai97.ps) from UAI-97, too. : Finally, the Introduction to Bayesian Inference and BBNs(http://www.cs.ust.hk/~samee/bayesian/intro.html) is a good, general meta-tutorial, with pointers to several recent AAAI tutorials (including[Friedman and Goldszmidt, 1998] on learning BBNs from data). The links are dead, though. Here is an archive of the tutorial on Nir's Stanford page: : http://robotics.Stanford.EDU/people/nir/tutorial/index.html : I also recommend the Breese and Koller general tutorial on BBNs: : http://www.research.microsoft.com/users/breese/tutorial/ : (Daphne's tutorial link at AAAI is dead, but the slides on her Stanford page [2-up and 6-up] are still there: : http://robotics.stanford.edu/~koller/BNtut/tut2.ps : http://robotics.stanford.edu/~koller/BNtut/tut6.ps : I guess the Stanford webmasters are just more reliable. :-)) : Hope this helps, : Bill

Optional Link URL:
Link Title:
Optional Image URL: