[Pellet-users] Open world reasoning with quantifiied restrictions

[Ron Alford]

John M. Linebarger wrote:
>
> I think I understand that, but I have two additional questions.  First,
> is there another way to approach this problem in OWL such that I achieve
> the result that I want?  And second, what is the purpose of quantified
> restrictions in the first place?  It would seem that all such quantified
> restrictions would run afoul of the open world assumption, such that
> they would be useless from a reasoning point of view.

Hi John,
If you know the number of connotations that a word has, you can state it:

<Word rdf:ID="....">
 <rdf:type>
  <owl:Restriction>
   <owl:onProperty rdf:about="...#hasConnotation"/>
   <owl:cardinality rdf:datatype="...">3</owl:cardinality>
  </owl:Restriction>
 </rdf:type>
 <hasPositive....
 <hasNegative...
 <hasNegative...
</Word>

The downside to this is that you won't be able to aggregate information
about a word (since you just specified that you knew every connotation).
 One way to get around this might be to use definition sets:
<Word ...>
  <definitionSet>
    <Definition>
      <rdf:type> /* 3 conntations
         ....
      </rdf:type>
      <hasPositive...
      ...
    </Definitiion>
  </definitionSet>
</Word>

Now you can say that by /your/ definition, this is or is not a dirty
word.  Then you can classify words by "SomeoneSaysItsDirty subClassOf
Exists on property definitionSet a definition with a negative connotation".

Then you could also specify disputed words, and possibly-polite words.
However, since it's open world, you still can't know if things are
definitely polite or impolite (unless you just go ahead and state it).
I think there's an amusing example of this between spanish in Spain and
spanish in the rest of the world, but I don't know enough to remember it.


But as to cardinality in general...

Both min and max cardinality can be used in classification and querying
(a classic being "find me all parents with more|less than 3 children").

Max cardinality is often used to force equality (FPs IFPs are basically
sugar for this).  I don't use minCard much (usually exists is good enough).

There may be other uses (certainly other people are more qualified to
say), but these are what I've seen.


-Ron

-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 251 bytes
Desc: OpenPGP digital signature
Url : http://lists.owldl.com/pipermail/pellet-users/attachments/20070328/ae290483/attachment.pgp