Does anybody know of downloadable lisp code to perform Bayesian
probabilistic reasoning? All the packages I have found lack
documentation and seem to be non-straightforward to use. To start with I
would like to transfer a simple probabilistic inference network, which I
previously built in MSNBX, to lisp.
Thanks
Juergen
On Thu, 20 Feb 2003 16:19:15 -0800, Juergen Meyer
<·······@u.washington.edu> wrote:
>Does anybody know of downloadable lisp code to perform Bayesian
>probabilistic reasoning? All the packages I have found lack
>documentation and seem to be non-straightforward to use. To start with I
>would like to transfer a simple probabilistic inference network, which I
>previously built in MSNBX, to lisp.
>
>Thanks
>Juergen
>
you could try xlisp-stat, which has some bayesian packages.
Marc
--
------------------------------------------------------------------------------
email: marc dot hoffmann at users dot whh dot wau dot nl
------------------------------------------------------------------------------
Hi Juergen,
if you are looking for software on Belief Networks, there's a listing on
http://www.stat.washington.edu/almond/belief.html. There are quite a
number of Common Lisp implementations listed, but I haven't checked if they
are usable (i.e., documented and still supported by someone).
Matthias
Juergen Meyer wrote:
> Does anybody know of downloadable lisp code to perform Bayesian
> probabilistic reasoning? All the packages I have found lack
> documentation and seem to be non-straightforward to use. To start with I
> would like to transfer a simple probabilistic inference network, which I
> previously built in MSNBX, to lisp.
>
> Thanks
> Juergen
"Juergen Meyer" <·······@u.washington.edu> schrieb im Newsbeitrag
·····················@u.washington.edu...
> Does anybody know of downloadable lisp code to perform Bayesian
> probabilistic reasoning? All the packages I have found lack
> documentation and seem to be non-straightforward to use. To start with I
> would like to transfer a simple probabilistic inference network, which I
> previously built in MSNBX, to lisp.
>
> Thanks
> Juergen
>
I have developed a CL program for bayesian networks some years ago.
It offers a graphical user interface for definition and testing bayesian
networks.
It uses the efficient Hugin propagation algorithm. The graphical user
interface is implemented,
however, only for allegro 3.0.1. The core is ANSI-CL and works with
Lispworks 4.2.7.
The program BayNet supports two source formats: Net and Lisp.
Let me know, if you want to have it.
Example
1. Lisp-Syntax
(defnet wetgrass)
(defnode regen (ja nein))
(defnode sprinkler (ja nein))
(defnode watson (ja nein))
(defnode holmes (ja nein))
(defpotential (regen ! )
#(0.2 0.8))
(defpotential (sprinkler ! )
#(0.1 0.9))
(defpotential (watson ! regen)
#2A((1.0 0.2) (0.0 0.8)))
(defpotential (holmes ! sprinkler regen)
#3A(((1.0 0.9) (1.0 0.0))
((0.0 0.1) (0.0 1.0))))
2. NET-Syntax
net
{
node_size = (80 40);
HRUNTIME_Monitor_InitStates = "5";
...
}
node Holmes
{
label = "Holmes?";
position = (140 0);
states = ("yes" "no");
}
node Watson
{
label = "Watson?";
states = ("yes" "no");
}
node Sprinkler
{
label = "Sprinkler?";
position = (220 70);
states = ("yes" "no");
}
node Rain
{
label = "Rain?";
position = (80 70);
states = ("yes" "no");
}
potential (Holmes | Sprinkler Rain)
{
data = ((( 1 0 ) % yes yes
( 0.9 0.1 )) % yes no
(( 1 0 ) % no yes
( 0 1 ))); % no no
}
potential (Watson | Rain)
{
data = (( 1 0 ) % yes
( 0.2 0.8 )); % no
}
potential (Sprinkler |)
{
data = ( 0.1 0.9 );
}
potential (Rain |)
{
data = ( 0.2 0.8 );
}
Benno
············@berlin.de
"Benno Biewer" <············@berlin.de> wrote in message news:<···············@news.t-online.com>...
> "Juergen Meyer" <·······@u.washington.edu> schrieb im Newsbeitrag
> ·····················@u.washington.edu...
> > Does anybody know of downloadable lisp code to perform Bayesian
> > probabilistic reasoning? All the packages I have found lack
...
> Let me know, if you want to have it.
Might it be possible, then, for you to contribute this code to one of
the CL repositories?
Regards,
Olivier
There is obviously more interest than I thought. Now, I itend to contribute
this code and (eventually further code ...) to one of the CL repositories.
Benno
"Olivier Drolet" <·······@mac.com> schrieb im Newsbeitrag
·································@posting.google.com...
> "Benno Biewer" <············@berlin.de> wrote in message
news:<···············@news.t-online.com>...
> > "Juergen Meyer" <·······@u.washington.edu> schrieb im Newsbeitrag
> > ·····················@u.washington.edu...
> > > Does anybody know of downloadable lisp code to perform Bayesian
> > > probabilistic reasoning? All the packages I have found lack
> ...
> > Let me know, if you want to have it.
>
> Might it be possible, then, for you to contribute this code to one of
> the CL repositories?
>
> Regards,
>
> Olivier
Benno's code helped me a lot. I didn't need the GUI and just had to make
minor modifications in order to run it on Allegro lisp on unix. The
entire code is just a couple of files compared to other packages which
come with tons of files but no documentation. The code does exactly what
I needed which is to build a straightforward Bayesian network. Thanks
again, Benno, for your help!
Juergen
Juergen Meyer wrote:
> Does anybody know of downloadable lisp code to perform Bayesian
> probabilistic reasoning? All the packages I have found lack
> documentation and seem to be non-straightforward to use. To start with I
> would like to transfer a simple probabilistic inference network, which I
> previously built in MSNBX, to lisp.
>
> Thanks
> Juergen
>
--
Juergen Meyer, Ph.D.
Senior Fellow
Radiation Oncology
Division of Radiation Physics
University of Washington Medical Center
Box 356043
Seattle, WA 98195-6043, USA
Phone: 206-598-6671