Can anyone direct me to research on structure learning of dags based on sampling techniques? Either using the Cooper and Herskovits or the Heckerman et al BDe/BGe metrics, or others? I am aware of the group at Monash using Minimum Message Length, and of 1994 work by Lam and Bacchus using MDL, but have not found any recently published work on search and score based structure learning apart from Giudici and Green's 1999 work for decomposable undirected graphs.
Thanks in advance for any feedback. Helen. -- ------------------------------------ Helen Armstrong [EMAIL PROTECTED] ph: + 61 2 9385 6904 fax: 9385 7123 School of Mathematics UNSW SYDNEY NSW 2052