[SciPy-user] is there any plan to import BNT(bayesian network toolkit) from matlab?

[Hoyt Koepke]

Very good news; I'll be following the development closely.

As for references, one suggestion is Kevin Murphy's PhD thesis on bayes nets.

http://www.cs.ubc.ca/~murphyk/Thesis/thesis.html

Just a thought.

--Hoyt

On Thu, Apr 10, 2008 at 11:33 AM, Stéfan van der Walt <stefan at sun.ac.za> wrote:
> Hi,
>
>  We have a master's student who is interested in doing the initial
>  port, and I hope we can form a productive collaboration: he has the
>  time and you guys have a vision.  It would be helpful if you could
>  either discuss the idea further here, or expand the wiki page, so that
>  I can give him a concrete plan to start with.
>
>  Amongs other things, I'd like to know which parts are most important
>  to port (i.e. which parts are not already provided by other
>  libraries), how we should refactor the interface to best make use of
>  Python's strong object oriented facilities (maybe this will develop
>  intuitively), and which references to use in docstrings (Chris
>  Bishop's book looks like a good place to start?).
>
>  Regards
>  Stéfan
>
>
>
>  On 08/04/2008, Karl Young <Karl.Young at ucsf.edu> wrote:
>  >
>  >  Just confirming Andrew's assessment of my situation (the description of
>  >  his situation fits mine as well). I'm certainly willing to put some time
>  >  into this project but don't have the time to organize it; while not an
>  >  expert in either Bayes nets or python hacking I've done a fair amount of
>  >  work with both (as well as Matlab in the ancient past, which might be of
>  >  some use with the port).
>  >
>  >
>  >  >See http://scipy.org/scipy/scikits/wiki/BayesNet . There are plans, the
>  >  >BNT license is scipy-compatible, and there are already 5 people
>  >  >interesting in developing. Speaking for myself, however, this has taken
>  >  >a lower priority than I initially expected due to the ever-changing
>  >  >energy landscape of a research program -- I'm simply too busy on other
>  >  >stuff right now to contribute meaningfully. I suspect the situation may
>  >  >be similar for the others listed.
>  >  >
>  >  >Nonetheless, we would welcome effort in this direction.
>  >  >
>  >  >-Andrew
>  _______________________________________________
>  SciPy-user mailing list
>  SciPy-user at scipy.org
>  http://projects.scipy.org/mailman/listinfo/scipy-user
>



-- 
+++++++++++++++++++++++++++++++++++
Hoyt Koepke
UBC Department of Computer Science
http://www.cs.ubc.ca/~hoytak/
hoytak at gmail.com
+++++++++++++++++++++++++++++++++++