Sounds like one for the Pellet folks (depending on what the error is that you get).

Though if your data is static then you could materialize the Pellet results (or the entire deductive closure, if there is one) in a plain model and then instantiate the rule-based model on top of that.

May not be a very usable architecture in general though and certainly little us if your data is not static.

Dave

On 02/12/13 15:06, Miguel Bento Alves wrote:
Hi,

It was what I tried. Doesn¹t work, Pellet returns error.

MBA

On 02/12/13 14:26, "Dave Reynolds" <[email protected]> wrote:

Hi,

I don't have access to Pellet so can't construct a complete example, but
all I meant was something like:

     Model myPelletModel = ...
     GenericRuleReasoner reasoner = new GenericRuleReasoner(rules);
     InfModel inf = ModelFactory.createInfModel(reasoner, myPelletModel);

Dave

On 02/12/13 13:57, Miguel Bento Alves wrote:
Dear Dave,

Thanks for your answer. Can you give a short example how I can put a
jena
rule-based inference model on top of a Pellet-backed model? I tried to
combine both but I didn't had success.

Miguel


On Mon, Dec 2, 2013 at 12:14 PM, Dave Reynolds
<[email protected]>wrote:

On 02/12/13 11:23, Miguel Bento Alves wrote:

Hi there,

Jena does not provide OWL2, correct?


Correct.


   My problem is that I want to use OWL2 inference, specifically to the
qualified cardinalities inference (for instance, and is just an
example,
my class IronMan is defined as all sportsman that plays at least 3
sports). With Pellet I can do OWL2 inference but I can't combine with
jena
rules. In my problem, jena rules are useful because I can develop
built-in
functions.

Some questions:

There are somehow to combine OWL2 inference with jena rules?


If your data is static then you may be able to put a jena rule-based
inference model on top of a Pellet-backed model.


   Do you know somehow to implement qualified cardinalities in jena?


I wouldn't recommend trying that. The lack of unique name assumption
means
that counting how many distinct things you have or can infer is very
tricky
in a simple rule based system.

Dave







Reply via email to