a question on property

classic Classic list List threaded Threaded
9 messages Options
Reply | Threaded
Open this post in threaded view
|

a question on property

Y.W.Yang
Hi,
 
DL with concrete domain can be used to express formalism like "a has visited b 3 times". In OWL, we can bulid class A and B, for instances a and b, a property with domain (A) and range (B), but how to express "3 times"?
 
Thanks,
Yanwu
Reply | Threaded
Open this post in threaded view
|

Re: a question on property

Y.W.Yang
Hello Hai,

Thanks for your reply. Yes I used QCR in building ontology...

In addition, there is one case confused me. I don't want to give detailed cardinality restriction between two classes (e.g. A and B), since the value of relationships are different to every pair of (a1, b1), (a2,b2). For example, I am intended to express the facts "a1 visited b1 4 times", and "a2 visited b2", that is, how to define the values of relationships between instances? (instead of classes).  

Best regards,
Yanwu
 
 
======= At 2006-01-26, 19:36:51 you wrote: =======

>Hi Yanwu,
>I am not very sure if I understood what you mean by "3 times".  Is  
>Qualified cardinality restrictions (QCR) (http://www.cs.vu.nl/~guus/ 
>public/qcr.html) what you want?
>If so, current version of OWL DL does not support QCR yet.  Protege-
>OWL supports QCRs indirectly by using a specially interpreted  
>property on cardinality restrictions.
>Please let et me know if I misunderstood your question.
>
>Regards
>Hai
>
>
>On 26 Jan 2006, at 04:14, Y.W.Yang wrote:
>
>> Hi,
>>
>> DL with concrete domain can be used to express formalism like "a  
>> has visited b 3 times". In OWL, we can bulid class A and B, for  
>> instances a and b, a property with domain (A) and range (B), but  
>> how to express "3 times"?
>>
>> Thanks,
>> Yanwu
>
>Regards
>Hai
>
>
>Research Associate
>Department of Computer Science
>Kilburn Building
>University of Manchester
>Oxford Road
>Manchester M13 9PL
>UK
>Homepage: http://www.cs.man.ac.uk/~hwang
>
>
>
>-------------------------------------------------------------------------
>To unsubscribe go to http://protege.stanford.edu/community/subscribe.html
>
>
>
>.

= = = = = = = = = = = = = = = = = = = =
                       
Y.W.Yang
[hidden email]
2006-01-26





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

Reply | Threaded
Open this post in threaded view
|

Re: a question on property

Hai Wang
In reply to this post by Y.W.Yang
Hi Yanwu,
I am not very sure if I understood what you mean by "3 times".  Is  
Qualified cardinality restrictions (QCR) (http://www.cs.vu.nl/~guus/ 
public/qcr.html) what you want?
If so, current version of OWL DL does not support QCR yet.  Protege-
OWL supports QCRs indirectly by using a specially interpreted  
property on cardinality restrictions.
Please let et me know if I misunderstood your question.

Regards
Hai


On 26 Jan 2006, at 04:14, Y.W.Yang wrote:

> Hi,
>
> DL with concrete domain can be used to express formalism like "a  
> has visited b 3 times". In OWL, we can bulid class A and B, for  
> instances a and b, a property with domain (A) and range (B), but  
> how to express "3 times"?
>
> Thanks,
> Yanwu

Regards
Hai


Research Associate
Department of Computer Science
Kilburn Building
University of Manchester
Oxford Road
Manchester M13 9PL
UK
Homepage: http://www.cs.man.ac.uk/~hwang



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

Reply | Threaded
Open this post in threaded view
|

Re: a question on property

Y.W.Yang
In reply to this post by Y.W.Yang
Hello protege-owl,


Hi, is there someone help with the following question?

There is one case confused me. I don't want to give detailed cardinality restriction between two classes (e.g. A and B), since the value of relationships are different to every pair of (a1, b1), (a2,b2). For example, I am intended to express the facts "a1 visited b1 4 times", and "a2 visited b2", that is, how to define the values of relationships between instances? (instead of classes).  

Best regards,
Yanwu

>
>  
>======= At 2006-01-26, 19:36:51 you wrote: =======
>
>>Hi Yanwu,
>>I am not very sure if I understood what you mean by "3 times".  Is  
>>Qualified cardinality restrictions (QCR) (http://www.cs.vu.nl/~guus/ 
>>public/qcr.html) what you want?
>>If so, current version of OWL DL does not support QCR yet.  Protege-
>>OWL supports QCRs indirectly by using a specially interpreted  
>>property on cardinality restrictions.
>>Please let et me know if I misunderstood your question.
>>
>>Regards
>>Hai
>>
>>
>>On 26 Jan 2006, at 04:14, Y.W.Yang wrote:
>>
>>> Hi,
>>>
>>> DL with concrete domain can be used to express formalism like "a  
>>> has visited b 3 times". In OWL, we can bulid class A and B, for  
>>> instances a and b, a property with domain (A) and range (B), but  
>>> how to express "3 times"?
>>>
>>> Thanks,
>>> Yanwu
>>
>>Regards
>>Hai
>>
>>
>>Research Associate
>>Department of Computer Science
>>Kilburn Building
>>University of Manchester
>>Oxford Road
>>Manchester M13 9PL
>>UK
>>Homepage: http://www.cs.man.ac.uk/~hwang
>>
>>
>>
>>-------------------------------------------------------------------------
>>To unsubscribe go to http://protege.stanford.edu/community/subscribe.html
>>
>>
>>
>>.
>
>= = = = = = = = = = = = = = = = = = = =
>
>Y.W.Yang
>[hidden email]
>2006-01-26
>

= = = = = = = = = = = = = = = = = = = =
 
Y.W.Yang
[hidden email]
2006-01-27




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

Reply | Threaded
Open this post in threaded view
|

Re: a question on property

Y.W.Yang
In reply to this post by Y.W.Yang
Hello, Hai,
 
Thanks a lot. Actually I try to find a way to attach values (not cardinality) to relationships between concepts, or between individuals. If we model relationships as a class, then it's independent, then somewhat some information on relationships is lost. The fact I want model is "a has visit b1 twice, b2 5 times, b3 3 times, etc...". Ok, the approach of modelling relationships as classes can work in this case. But it's yet not good enough...
 
Thanks again,
Yanwu
 

Hi Yanwu,
I looked your problem again. I maybe misunderstood you question before. It seems that you want to express a property involving three elements, i.e Visisting (From, To, Times). If that was the case, QCR may not be suitable.
In OWL, all properties are  binary relations used to link two individuals or an individual and a value. If you want to use relations to link an individual to more than just one individual or value, you have to reify the relation to a class first.
(http://smi-web.stanford.edu/people/noy/nAryRelations/n-aryRelations-2nd-WD.html for the details). In you case, you should define a class named Visiting, three properties, visitFrom, visitTo and visitTimes.
 
Is that what you want to model?
 
Regards
Hai
 
 


 
On 1/26/06, Y.W.Yang <[hidden email]> wrote:
Hello Hai,

Thanks for your reply. Yes I used QCR in building ontology...

In addition, there is one case confused me. I don't want to give detailed cardinality restriction between two classes (e.g. A and B), since the value of relationships are different to every pair of (a1, b1), (a2,b2). For example, I am intended to express the facts "a1 visited b1 4 times", and "a2 visited b2", that is, how to define the values of relationships between instances? (instead of classes).

Best regards,
Yanwu


======= At 2006-01-26, 19:36:51 you wrote: =======

>Hi Yanwu,
>I am not very sure if I understood what you mean by "3 times".  Is
>Qualified cardinality restrictions (QCR) (http://www.cs.vu.nl/~guus/
>public/qcr.html) what you want?
>If so, current version of OWL DL does not support QCR yet.  Protege-
>OWL supports QCRs indirectly by using a specially interpreted
>property on cardinality restrictions.
>Please let et me know if I misunderstood your question.
>
>Regards
>Hai
>
>
>On 26 Jan 2006, at 04:14, Y.W.Yang wrote:
>
>> Hi,
>>
>> DL with concrete domain can be used to express formalism like "a
>> has visited b 3 times". In OWL, we can bulid class A and B, for
>> instances a and b, a property with domain (A) and range (B), but
>> how to express "3 times"?
>>
>> Thanks,
>> Yanwu
>
>Regards
>Hai
>
>
>Research Associate
>Department of Computer Science
>Kilburn Building
>University of Manchester
>Oxford Road
>Manchester M13 9PL
>UK
>Homepage: http://www.cs.man.ac.uk/~hwang
>
>
>
>-------------------------------------------------------------------------
>To unsubscribe go to http://protege.stanford.edu/community/subscribe.html
>
>
>
>.

= = = = = = = = = = = = = = = = = = = =

Y.W.Yang
[hidden email]
2006-01-26





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



Reply | Threaded
Open this post in threaded view
|

Re: a question on property

Y.W.Yang
In reply to this post by Y.W.Yang
Hello Maurizio,

Thanks for your information. I ever read your papers on DLR and others on DL with concrete domain extensions. At DL level it's possible to present these kinds of facts. However at the moment of implement I tried to use Protege to construct our previous DL-based ontology. Then there seems a problem to deal with. The only approach left is to model these n-ary relationships as classes.  
                                   
Best regards,
Yanwu
 
======= At 2006-01-27, 20:27:28 you wrote: =======

>Y.W.Yang wrote:
>
>> Hello, Hai,
>>  
>> Thanks a lot. Actually I try to find a way to attach values (not
>> cardinality) to relationships between concepts, or between individuals.
>> If we model relationships as a class, then it's independent, then
>> somewhat some information on relationships is lost. The fact I want
>> model is "a has visit b1 twice, b2 5 times, b3 3 times, etc...". Ok, the
>> approach of modelling relationships as classes can work in this case.
>> But it's yet not good enough...
>>  
>> Thanks again,
>> Yanwu
>>  
>
>I agree: the problem of specifying properties of relationships is a genuine
>modeling problem. From a language point of view, you might be interested in
>looking at DLR, a Description Logics providing specific modeling constructs for
>n-ary relations. Since DLR allows you to specify identification and functional
>constraints on n-ary relations, attributes of relationships can be easily
>modeled in this logic.
>
>Ref.:
>Diego Calvanese, Giuseppe De Giacomo, and Maurizio Lenzerini. "Identification
>constraints and functional dependencies in description logics". In Proc. of the
>17th Int. Joint Conf. on Artificial Intelligence (IJCAI 2001), pages 155-160, 2001.
>
>Best.
>
>--
>Maurizio
>
>
>------------------------------------------------------------
>Prof. Maurizio Lenzerini
>Dipartimento di Informatica e Sistemistica "Antonio Ruberti"
>Università degli Studi di Roma "La Sapienza"
>Via Salaria 113,  I-00198  Roma,  Italy
>Tel:       +39 - 06 - 8841954
>Fax:       +39 - 06 - 85300849
>E-mail:    [hidden email]
>Home page: http://www.dis.uniroma1.it/~lenzerini
>
>-------------------------------------------------------------------------
>To unsubscribe go to http://protege.stanford.edu/community/subscribe.html

= = = = = = = = = = = = = = = = = = = =
                       
Y.W.Yang
[hidden email]
2006-01-27

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

Reply | Threaded
Open this post in threaded view
|

Re: a question on property

Hai Wang-2
In reply to this post by Y.W.Yang
Hi Yanwu,
I looked your problem again. I maybe misunderstood you question before. It seems that you want to express a property involving three elements, i.e Visisting (From, To, Times). If that was the case, QCR may not be suitable.
In OWL, all properties are  binary relations used to link two individuals or an individual and a value. If you want to use relations to link an individual to more than just one individual or value, you have to reify the relation to a class first.
(http://smi-web.stanford.edu/people/noy/nAryRelations/n-aryRelations-2nd-WD.html for the details). In you case, you should define a class named Visiting, three properties, visitFrom, visitTo and visitTimes.
 
Is that what you want to model?
 
Regards
Hai
 
 


 
On 1/26/06, Y.W.Yang <[hidden email]> wrote:
Hello Hai,

Thanks for your reply. Yes I used QCR in building ontology...

In addition, there is one case confused me. I don't want to give detailed cardinality restriction between two classes (e.g. A and B), since the value of relationships are different to every pair of (a1, b1), (a2,b2). For example, I am intended to express the facts "a1 visited b1 4 times", and "a2 visited b2", that is, how to define the values of relationships between instances? (instead of classes).

Best regards,
Yanwu


======= At 2006-01-26, 19:36:51 you wrote: =======

>Hi Yanwu,
>I am not very sure if I understood what you mean by "3 times".  Is
>Qualified cardinality restrictions (QCR) (http://www.cs.vu.nl/~guus/
>public/qcr.html) what you want?
>If so, current version of OWL DL does not support QCR yet.  Protege-
>OWL supports QCRs indirectly by using a specially interpreted
>property on cardinality restrictions.
>Please let et me know if I misunderstood your question.
>
>Regards
>Hai
>
>
>On 26 Jan 2006, at 04:14, Y.W.Yang wrote:
>
>> Hi,
>>
>> DL with concrete domain can be used to express formalism like "a
>> has visited b 3 times". In OWL, we can bulid class A and B, for
>> instances a and b, a property with domain (A) and range (B), but
>> how to express "3 times"?
>>
>> Thanks,
>> Yanwu
>
>Regards
>Hai
>
>
>Research Associate
>Department of Computer Science
>Kilburn Building
>University of Manchester
>Oxford Road
>Manchester M13 9PL
>UK
>Homepage: http://www.cs.man.ac.uk/~hwang
>
>
>
>-------------------------------------------------------------------------
>To unsubscribe go to http://protege.stanford.edu/community/subscribe.html
>
>
>
>.

= = = = = = = = = = = = = = = = = = = =

Y.W.Yang
[hidden email]
2006-01-26





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


Reply | Threaded
Open this post in threaded view
|

Re: a question on property

Maurizio Lenzerini
In reply to this post by Y.W.Yang
Y.W.Yang wrote:

> Hello, Hai,
>  
> Thanks a lot. Actually I try to find a way to attach values (not
> cardinality) to relationships between concepts, or between individuals.
> If we model relationships as a class, then it's independent, then
> somewhat some information on relationships is lost. The fact I want
> model is "a has visit b1 twice, b2 5 times, b3 3 times, etc...". Ok, the
> approach of modelling relationships as classes can work in this case.
> But it's yet not good enough...
>  
> Thanks again,
> Yanwu
>  

I agree: the problem of specifying properties of relationships is a genuine
modeling problem. From a language point of view, you might be interested in
looking at DLR, a Description Logics providing specific modeling constructs for
n-ary relations. Since DLR allows you to specify identification and functional
constraints on n-ary relations, attributes of relationships can be easily
modeled in this logic.

Ref.:
Diego Calvanese, Giuseppe De Giacomo, and Maurizio Lenzerini. "Identification
constraints and functional dependencies in description logics". In Proc. of the
17th Int. Joint Conf. on Artificial Intelligence (IJCAI 2001), pages 155-160, 2001.

Best.

--
Maurizio


------------------------------------------------------------
Prof. Maurizio Lenzerini
Dipartimento di Informatica e Sistemistica "Antonio Ruberti"
Università degli Studi di Roma "La Sapienza"
Via Salaria 113,  I-00198  Roma,  Italy
Tel:       +39 - 06 - 8841954
Fax:       +39 - 06 - 85300849
E-mail:    [hidden email]
Home page: http://www.dis.uniroma1.it/~lenzerini

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

Reply | Threaded
Open this post in threaded view
|

Re: a question on property

Wei Xing
In reply to this post by Hai Wang-2
Hi,

I think Hai's approach is correct.

Take a look:

http://www.w3.org/TR/swbp-n-aryRelations/

Regards

Wei

Hai Wang wrote:

> Hi Yanwu,
> I looked your problem again. I maybe misunderstood you question
> before. It seems that you want to express a property involving three
> elements, i.e Visisting (From, To, Times). If that was the case, QCR
> may not be suitable.
> In OWL, all properties are /binary/ relations used to link two
> individuals or an individual and a value. If you want to use relations
> to link an individual to more than just one individual or value, you
> have to reify the relation to a class first.
> (http://smi-web.stanford.edu/people/noy/nAryRelations/n-aryRelations-2nd-WD.html
> for the details). In you case, you should define a class named
> Visiting, three properties, visitFrom, visitTo and visitTimes.
> Is that what you want to model?
> Regards
> Hai
>
>
> On 1/26/06, *Y.W.Yang* <[hidden email]> wrote:
>
>     Hello Hai,
>
>     Thanks for your reply. Yes I used QCR in building ontology...
>
>     In addition, there is one case confused me. I don't want to give
>     detailed cardinality restriction between two classes (e.g. A and
>     B), since the value of relationships are different to every pair
>     of (a1, b1), (a2,b2). For example, I am intended to express the
>     facts "a1 visited b1 4 times", and "a2 visited b2", that is, how
>     to define the values of relationships between instances? (instead
>     of classes).
>
>     Best regards,
>     Yanwu
>
>
>     ======= At 2006-01-26, 19:36:51 you wrote: =======
>
>     >Hi Yanwu,
>     >I am not very sure if I understood what you mean by "3 times". Is
>     >Qualified cardinality restrictions (QCR) (http://www.cs.vu.nl/~guus/
>     >public/qcr.html) what you want?
>     >If so, current version of OWL DL does not support QCR yet. Protege-
>     >OWL supports QCRs indirectly by using a specially interpreted
>     >property on cardinality restrictions.
>     >Please let et me know if I misunderstood your question.
>     >
>     >Regards
>     >Hai
>     >
>     >
>     >On 26 Jan 2006, at 04:14, Y.W.Yang wrote:
>     >
>     >> Hi,
>     >>
>     >> DL with concrete domain can be used to express formalism like "a
>     >> has visited b 3 times". In OWL, we can bulid class A and B, for
>     >> instances a and b, a property with domain (A) and range (B), but
>     >> how to express "3 times"?
>     >>
>     >> Thanks,
>     >> Yanwu
>     >
>     >Regards
>     >Hai
>     >
>     >
>     >Research Associate
>     >Department of Computer Science
>     >Kilburn Building
>     >University of Manchester
>     >Oxford Road
>     >Manchester M13 9PL
>     >UK
>     >Homepage: http://www.cs.man.ac.uk/~hwang
>     >
>     >
>     >
>     >-------------------------------------------------------------------------
>     >To unsubscribe go to
>     http://protege.stanford.edu/community/subscribe.html
>     >
>     >
>     >
>     >.
>
>     = = = = = = = = = = = = = = = = = = = =
>
>     Y.W.Yang
>     [hidden email]
>     2006-01-26
>
>
>
>
>
>     -------------------------------------------------------------------------
>     To unsubscribe go to
>     http://protege.stanford.edu/community/subscribe.html
>
>


--
============================================================
Wei Xing, M.Sc.
Research Associate                    Tel: 00357-22892663
Dept. of Computer Science             Fax: 00357-22892701
University of Cyprus                  email: [hidden email]
PO Box 20537
CY1678, Nicosia, CYPRUS


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