Hi David,
of course, there are many of them! If you mean by a probabilistic
model a model of conditional independence structure (that is,
statistical model defined by conditional independence restrictions)
then the number of non-graphical models increases exponentially with
the number of general graphical models. I am going to send you a
reference to a claim which implies this.
On the other hand, you may regard most of these probabilistic
models of conditional independence model as 'ugly' statistical
models because you are not able to find neat parametization
like in case of models determined by Bayesian networks.
Regards from
Milan Studeny
On Fri, 20 Jun 2003, David Bellot wrote:
> Hi,
>
> does it exist probabilistic models which are neither a bayesian network,
> nor a more general graphical model ?
>
> Thanks,
> David
>
> --
> ---------------------------------------------------
