Represent sentences in OWL

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

Represent sentences in OWL

Ans
I want to create sentences in OWL such as " I ate apple yesterday"
How can I create classes and properties for this?
Should I take all these as concepts?
Reply | Threaded
Open this post in threaded view
|

Re: Represent sentences in OWL

Martin Kuba
Dne 27.4.2012 13:46, Ans napsal(a):
> I want to create sentences in OWL such as " I ate apple yesterday"
> How can I create classes and properties for this?
> Should I take all these as concepts?

This is an interesting question. I think that expressing that sentence
in OWL (or in Description Logic) has two separate problems.

The first problem is how to express "I" and "yesterday". I think that
you cannot express that in OWL, because OWL describes ontologies which are
explicit *shared* descriptions of conceptualisations.
The meaning of "I" and "yesterday" is not shared, it depends on
who is saying it and when.

If you take a similar sentence "John ate apple on April 26, 2012",
the second problem is that OWL allows only binary relations,
and here you need a ternary relation

  eat(John,apple,"2012-04-26"^^xsd:date)

You can work around it by decomposing the ternary relation
by expressing eat as individual and using three properties, like

did(John,Eat)
what(Eat,apple)
when(Eat,"2012-04-26"^^xsd:date)

where "John" and "Eat" are individuals, "apple" is an individual from
a class of apples,  the date is a literal, "did" and "what" are object
properties and "when" is a data property.

Best regards

Martin
--
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
CERIT-SC                               Martin Kuba
Institute of Computer Science    email: [hidden email]
Masaryk University             http://www.ics.muni.cz/~makub/
Botanicka 68a, 60200 Brno, CZ     mobil: +420-603-533775
--------------------------------------------------------------
_______________________________________________
protege-owl mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-owl

Instructions for unsubscribing: http://protege.stanford.edu/doc/faq.html#01a.03
mhw
Reply | Threaded
Open this post in threaded view
|

Re: Represent sentences in OWL

mhw
It might be worth referencing the N-ary relations discussion from W3C (first hit on google).

BW,

Matt
Sent from my BlackBerry® wireless device

-----Original Message-----
From: Martin Kuba <[hidden email]>
Sender: [hidden email]
Date: Fri, 27 Apr 2012 15:09:44
To: User support for the Protege-OWL editor<[hidden email]>
Reply-To: User support for the Protege-OWL editor
        <[hidden email]>
Subject: Re: [protege-owl] Represent sentences in OWL

Dne 27.4.2012 13:46, Ans napsal(a):
> I want to create sentences in OWL such as " I ate apple yesterday"
> How can I create classes and properties for this?
> Should I take all these as concepts?

This is an interesting question. I think that expressing that sentence
in OWL (or in Description Logic) has two separate problems.

The first problem is how to express "I" and "yesterday". I think that
you cannot express that in OWL, because OWL describes ontologies which are
explicit *shared* descriptions of conceptualisations.
The meaning of "I" and "yesterday" is not shared, it depends on
who is saying it and when.

If you take a similar sentence "John ate apple on April 26, 2012",
the second problem is that OWL allows only binary relations,
and here you need a ternary relation

  eat(John,apple,"2012-04-26"^^xsd:date)

You can work around it by decomposing the ternary relation
by expressing eat as individual and using three properties, like

did(John,Eat)
what(Eat,apple)
when(Eat,"2012-04-26"^^xsd:date)

where "John" and "Eat" are individuals, "apple" is an individual from
a class of apples,  the date is a literal, "did" and "what" are object
properties and "when" is a data property.

Best regards

Martin
--
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
CERIT-SC                               Martin Kuba
Institute of Computer Science    email: [hidden email]
Masaryk University             http://www.ics.muni.cz/~makub/
Botanicka 68a, 60200 Brno, CZ     mobil: +420-603-533775
--------------------------------------------------------------
_______________________________________________
protege-owl mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-owl

Instructions for unsubscribing: http://protege.stanford.edu/doc/faq.html#01a.03
_______________________________________________
protege-owl mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-owl

Instructions for unsubscribing: http://protege.stanford.edu/doc/faq.html#01a.03
Reply | Threaded
Open this post in threaded view
|

Re: Represent sentences in OWL

Daniel Schober
In reply to this post by Martin Kuba
Dear Ans.

 Yes, as Martin says, the relative time ("yesterday") and Person ("I") indicators introduce difficulties. In general time models in owl tend to get very complex and they are practically to be regarded as an 'open issue'.
 I always saw the timeline outside of the Abox, i.e. the whole Abox specifies an instance of the world at one time point. Then you can iteratively manipulate the knowledge base (e.g. via rules of the form Present pattern-->Future pattern) to model a timeline (declarative present, imparative future). This makes it basically a simulation.
If you intend to exploit temporal reasoning a temporal logic like concurrent metatem, http://www.youtube.com/watch?v=ow2vbA9ohC0 might be useable.

After all, the way to model this in DL is dependent on what you intend to do with your model. The formal specific DL expressivity to be used is ultimately dictated by your intended application and the kind of inference you intend to do.

As DL is rooted in set theory, you usually need to model individuals & classes (sets of individuals).
"I" is an anaphora that refers to the individual "Ans Napsal" e.g. of a class Person.
"Apple" refers to an individual of the class Apple (the set of individuals with common properties that intrinsically define the Apple class).
If you would like to make a general ontological statement (always refering to whole sets of individuals) then you need to specify it in the T-box restrictions.
On the class level (T-box) in DL you need specify the quantification for the object property, i.e. if you want to describe the class
(the set of Persons) that only eats Apples (or nothing at all) you use the universal qantifyier i.e. in OWL Mancherster Syntax you state that
AppleEater eats ONLY Apple (not a complete closure)
If you intend to state that the each AppleEater always eats Apples and only Apples (nothing else) you state
AppleEater eats SOME Apple AND
AppleEater eats ONLY Apple (having the complete closure axiom)
If you intent to say that there is a set of Persons that eat among other things Apples then you use the existential quantifyer:
AppleEater eats SOME Apple
This means that for each and every Apple Eater there exists an individual Apple that is eaten by the AppleEater individual.

I would advice you to look into some DL tutorials like Baaders Description Logics Handbook to get the basics of DL semantics.

Regards,
 Daniel Schober.
 
----------------ursprüngliche Nachricht-----------------
Von: "Martin Kuba" [hidden email]
An: "User support for the Protege-OWL editor"
Datum: Fri, 27 Apr 2012 15:09:44 +0200
-------------------------------------------------
 
 

> Dne 27.4.2012 13:46, Ans napsal(a):
>> I want to create sentences in OWL such as " I ate apple yesterday"
>> How can I create classes and properties for this?
>> Should I take all these as concepts?
>
> This is an interesting question. I think that expressing that sentence
> in OWL (or in Description Logic) has two separate problems.
>
> The first problem is how to express "I" and "yesterday". I think that
> you cannot express that in OWL, because OWL describes ontologies which are
> explicit *shared* descriptions of conceptualisations.
> The meaning of "I" and "yesterday" is not shared, it depends on
> who is saying it and when.
>
> If you take a similar sentence "John ate apple on April 26, 2012",
> the second problem is that OWL allows only binary relations,
> and here you need a ternary relation
>
> eat(John,apple,"2012-04-26"^^xsd:date)
>
> You can work around it by decomposing the ternary relation
> by expressing eat as individual and using three properties, like
>
> did(John,Eat)
> what(Eat,apple)
> when(Eat,"2012-04-26"^^xsd:date)
>
> where "John" and "Eat" are individuals, "apple" is an individual from
> a class of apples, the date is a literal, "did" and "what" are object
> properties and "when" is a data property.
>
> Best regards
>
> Martin
> --
> ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> CERIT-SC Martin Kuba
> Institute of Computer Science email: [hidden email]
> Masaryk University http://www.ics.muni.cz/~makub/
> Botanicka 68a, 60200 Brno, CZ mobil: +420-603-533775
> --------------------------------------------------------------
> _______________________________________________
> protege-owl mailing list
> [hidden email]
> https://mailman.stanford.edu/mailman/listinfo/protege-owl
>
> Instructions for unsubscribing:
> http://protege.stanford.edu/doc/faq.html#01a.03
>
>

--

 

_______________________________________________
protege-owl mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-owl

Instructions for unsubscribing: http://protege.stanford.edu/doc/faq.html#01a.03
mhw
Reply | Threaded
Open this post in threaded view
|

Re: Represent sentences in OWL

mhw
Dear Daniel,

As an addition to your comments, I would specifically flag up all the existing work on temporal logics, and particularly the distinction between A and B mode temporal logics - wikipedia has a reasonable introduction.

These issues may well be more than the OP is interested in at present, but it is worth remembering that they exist.

BW,
Matt
Sent from my BlackBerry® wireless device

-----Original Message-----
From: Daniel Schober <[hidden email]>
Sender: [hidden email]
Date: Fri, 27 Apr 2012 17:17:29
To: User support for the Protege-OWL editor<[hidden email]>; User support for the Protege-OWL editor<[hidden email]>; Martin Kuba<[hidden email]>
Reply-To: User support for the Protege-OWL editor
        <[hidden email]>
Subject: Re: [protege-owl] Represent sentences in OWL

Dear Ans.

 Yes, as Martin says, the relative time ("yesterday") and Person ("I") indicators introduce difficulties. In general time models in owl tend to get very complex and they are practically to be regarded as an 'open issue'.
 I always saw the timeline outside of the Abox, i.e. the whole Abox specifies an instance of the world at one time point. Then you can iteratively manipulate the knowledge base (e.g. via rules of the form Present pattern-->Future pattern) to model a timeline (declarative present, imparative future). This makes it basically a simulation.
If you intend to exploit temporal reasoning a temporal logic like concurrent metatem, http://www.youtube.com/watch?v=ow2vbA9ohC0 might be useable.

After all, the way to model this in DL is dependent on what you intend to do with your model. The formal specific DL expressivity to be used is ultimately dictated by your intended application and the kind of inference you intend to do.

As DL is rooted in set theory, you usually need to model individuals & classes (sets of individuals).
"I" is an anaphora that refers to the individual "Ans Napsal" e.g. of a class Person.
"Apple" refers to an individual of the class Apple (the set of individuals with common properties that intrinsically define the Apple class).
If you would like to make a general ontological statement (always refering to whole sets of individuals) then you need to specify it in the T-box restrictions.
On the class level (T-box) in DL you need specify the quantification for the object property, i.e. if you want to describe the class
(the set of Persons) that only eats Apples (or nothing at all) you use the universal qantifyier i.e. in OWL Mancherster Syntax you state that
AppleEater eats ONLY Apple (not a complete closure)
If you intend to state that the each AppleEater always eats Apples and only Apples (nothing else) you state
AppleEater eats SOME Apple AND
AppleEater eats ONLY Apple (having the complete closure axiom)
If you intent to say that there is a set of Persons that eat among other things Apples then you use the existential quantifyer:
AppleEater eats SOME Apple
This means that for each and every Apple Eater there exists an individual Apple that is eaten by the AppleEater individual.

I would advice you to look into some DL tutorials like Baaders Description Logics Handbook to get the basics of DL semantics.

Regards,
 Daniel Schober.
 
----------------ursprüngliche Nachricht-----------------
Von: "Martin Kuba" [hidden email]
An: "User support for the Protege-OWL editor"
Datum: Fri, 27 Apr 2012 15:09:44 +0200
-------------------------------------------------
 
 

> Dne 27.4.2012 13:46, Ans napsal(a):
>> I want to create sentences in OWL such as " I ate apple yesterday"
>> How can I create classes and properties for this?
>> Should I take all these as concepts?
>
> This is an interesting question. I think that expressing that sentence
> in OWL (or in Description Logic) has two separate problems.
>
> The first problem is how to express "I" and "yesterday". I think that
> you cannot express that in OWL, because OWL describes ontologies which are
> explicit *shared* descriptions of conceptualisations.
> The meaning of "I" and "yesterday" is not shared, it depends on
> who is saying it and when.
>
> If you take a similar sentence "John ate apple on April 26, 2012",
> the second problem is that OWL allows only binary relations,
> and here you need a ternary relation
>
> eat(John,apple,"2012-04-26"^^xsd:date)
>
> You can work around it by decomposing the ternary relation
> by expressing eat as individual and using three properties, like
>
> did(John,Eat)
> what(Eat,apple)
> when(Eat,"2012-04-26"^^xsd:date)
>
> where "John" and "Eat" are individuals, "apple" is an individual from
> a class of apples, the date is a literal, "did" and "what" are object
> properties and "when" is a data property.
>
> Best regards
>
> Martin
> --
> ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> CERIT-SC Martin Kuba
> Institute of Computer Science email: [hidden email]
> Masaryk University http://www.ics.muni.cz/~makub/
> Botanicka 68a, 60200 Brno, CZ mobil: +420-603-533775
> --------------------------------------------------------------
> _______________________________________________
> protege-owl mailing list
> [hidden email]
> https://mailman.stanford.edu/mailman/listinfo/protege-owl
>
> Instructions for unsubscribing:
> http://protege.stanford.edu/doc/faq.html#01a.03
>
>

--

 

_______________________________________________
protege-owl mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-owl

Instructions for unsubscribing: http://protege.stanford.edu/doc/faq.html#01a.03
_______________________________________________
protege-owl mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-owl

Instructions for unsubscribing: http://protege.stanford.edu/doc/faq.html#01a.03
Ans
Reply | Threaded
Open this post in threaded view
|

Re: Represent sentences in OWL

Ans
In reply to this post by mhw
Thank you very much for everybody... I will try to do. At the beginning
I did not have any idea about it. Now I can understand.
Thank you