[protege-owl] is it possible to model user preferences as swrl atoms

classic Classic list List threaded Threaded
1 message Options
Reply | Threaded
Open this post in threaded view
|

[protege-owl] is it possible to model user preferences as swrl atoms

paolo-bussotti
Hello, Jayaram.
I am writing from another account but I'm the same Paolo whom
you answered recently.
Thanks for your answer:I would try to clarify what I really meant
in my request, as you asked me.

Im a novice about SWRL, but what I meant is this:

Suppose I have an ontology for tourism like the following:
   
   ............
   <owl:Class rdf:ID=NaturalPlaces/>
   <owl:Class rdf:ID=SeasidePlaces>
      <rdfs:subClassOf rdf:resource=#NaturalPlaces/>
   </owl:Class>
   <owl:Class rdf:ID=MountainsidePlaces>
      <rdfs:subClassOf rdf:resource=#NaturalPlaces/>
   </owl:Class>    
   <owl:Class rdf:ID=Tourist>
     <rdfs:subclassOf>
       <owl:Restriction>
          <owl:onProperty rdf:resource=#hasPreference/>
          <owl:someValuesFrom rdf:resource=#NaturalPlaces/>
       </owl:Restriction>        
   </owl:Class>
   <owl:ObjectProperty rdf:ID=hasPreference/>
   <SeasidePlaces rdf:ID="[specificSeasidePlace_1]"/>
   <SeasidePlaces rdf:ID="[specificSeasidePlace_2]"/>
   </rdf>
   
Now, if I import this ontology and create a tourist individual
from class Tourist, e.g.:

<ex:Tourist rdf:ID="myself">
  <ex:hasPreference rdf:resource:#ex:[specificSeasidePlace_1]/>
</ex:Tourist>

I'm forced to specify, for my preference, an individual of
SeasidePlace, (supposed I like the seaside), since I've defined
hasPreference as an ObjectProperty.

But if I wanted to express that I prefer the Seaside in general,
I cannot use that class as a value of preference and maintain the
capability of DL reasoning on the user preferences.

If, instead I used a SWRL rule to state the user preference,
for example:

              Tourist(?x)^hasPreference(?x,?y)=> NaturalPlace(?y)  

and then specify:
       Tourist(myself)^hasPreference(myself,?y)=>SeasidePlace(?y)    

this may allow rule-base reasoning on user preferences:
e.g. A query:
           Tourist(myself)^hasPreference(myself,?x)?
should give the answer:

    ?x=[specificSeasidePlace_1]
    ?x=[specificSeasidePlace_2]

MAYBE THIS IS NOT EXACTLY THE INTENDED MEANING OF
"GENERIC PREFERENCE"
(it implies to have preference to all instances)
but may be useful for reasoning on user-preference.

THE PROBLEM is that I've never worked with SWRL (and I'm asking you for a
good rule-engine, too), even if I'worked with Prolog ecc
but outside the Ontology framework.
DO YOU THINK THIS IDEA (I don't pretend itìs mine)COULD BE
A VALID MODELLING APPROACH?

Hoping to have cleared your doubts on my question,
I hope you can give me some advise.
Thanks a lot, anyway for your interest.

Paolo  




       

 

-------------------------------------------------------------------------
To unsubscribe go to http://protege.stanford.edu/community/subscribe.html