reasonning question

classic Classic list List threaded Threaded
22 messages Options
12
Reply | Threaded
Open this post in threaded view
|

reasonning question

Iulian Alecu
 
Hi,

I have the following situation for which I can't give my self an explanation
:
 
B partial A;
D partial C;

Restrictions R1 and R2 ;

CP1 complete (R1 some A) or (R2 some C);

CP2 complete (R1 some B) or (R2 some D);

At this point I agree that CP2 inferred_subclassof CP1;

Now, if I define
        CP3 complete (R1 some B) or (R2 some owl:Thing);

 I would expect to get after classification that both CP1 and CP2 are
subclasses of CP3, bu actually I get just CP2 as subclass of CP3.
 I feel that I'm missing an important point about reasonning in OWL but I
just can't figure out which.

Thanks ,

Iulian Alecu
-----------------------------------
Laboratoire SPIM
INSERM U729
-----------------------------------  

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

Reply | Threaded
Open this post in threaded view
|

Re: reasonning question

Hai Wang
Hi Lulian,

The reason is that the restriction (R1 some B) is not subsumed by  
the  restriction (R1 some A).
For example, suppose we have another class F
F partial A, F is disjoint with B.

(R1 some F) will be a subclass of (R1 some A) ( also subclass of  
CP1), but it is not subsumed by either (R1 SOME B) or  ( R2 some  
OWL:thing) which are the definition of CP3.


Regards
Hai

On 6 Mar 2006, at 14:52, Iulian Alecu wrote:

>
> Hi,
>
> I have the following situation for which I can't give my self an  
> explanation
> :
>
> B partial A;
> D partial C;
>
> Restrictions R1 and R2 ;
>
> CP1 complete (R1 some A) or (R2 some C);
>
> CP2 complete (R1 some B) or (R2 some D);
>
> At this point I agree that CP2 inferred_subclassof CP1;
>
> Now, if I define
> CP3 complete (R1 some B) or (R2 some owl:Thing);
>
>  I would expect to get after classification that both CP1 and CP2 are
> subclasses of CP3, bu actually I get just CP2 as subclass of CP3.
>  I feel that I'm missing an important point about reasonning in OWL  
> but I
> just can't figure out which.
>
> Thanks ,
>
> Iulian Alecu
> -----------------------------------
> Laboratoire SPIM
> INSERM U729
> -----------------------------------
>
> ----------------------------------------------------------------------
> ---
> To unsubscribe go to http://protege.stanford.edu/community/ 
> subscribe.html
>

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: reasonning question

John Goodwin
In reply to this post by Iulian Alecu
Try looking at the problem with proper examples, e.g.

Male partial Person
Cat partial Animal

CP1 = hasChild some Person and hasPet some Animal
CP2 = hasChild some Male and hasPet some Cat

CP3 = hasChild some Male and hasPet some owl:Thing


Why could you now expect CP1 to be a subclass of CP3?

John

> -----Original Message-----
> From: [hidden email]
> [mailto:[hidden email]] On Behalf Of
> Iulian Alecu
> Sent: 06 March 2006 14:53
> To: [hidden email]
> Subject: [protege-owl] reasonning question
>
>  
> Hi,
>
> I have the following situation for which I can't give my self
> an explanation
> :
>  
> B partial A;
> D partial C;
>
> Restrictions R1 and R2 ;
>
> CP1 complete (R1 some A) or (R2 some C);
>
> CP2 complete (R1 some B) or (R2 some D);
>
> At this point I agree that CP2 inferred_subclassof CP1;
>
> Now, if I define
> CP3 complete (R1 some B) or (R2 some owl:Thing);
>
>  I would expect to get after classification that both CP1 and
> CP2 are subclasses of CP3, bu actually I get just CP2 as
> subclass of CP3.
>  I feel that I'm missing an important point about reasonning
> in OWL but I just can't figure out which.
>
> Thanks ,
>
> Iulian Alecu
> -----------------------------------
> Laboratoire SPIM
> INSERM U729
> -----------------------------------  
>
> --------------------------------------------------------------
> -----------
> To unsubscribe go to
> http://protege.stanford.edu/community/subscribe.html
>
>
.


This email is only intended for the person to whom it is addressed and may contain confidential information. If you have received this email in error, please notify the sender and delete this email which must not be copied, distributed or disclosed to any other person.

Unless stated otherwise, the contents of this email are personal to the writer and do not represent the official view of Ordnance Survey. Nor can any contract be formed on Ordnance Survey's behalf via email. We reserve the right to monitor emails and attachments without prior notice.

Thank you for your cooperation.

Ordnance Survey
Romsey Road
Southampton SO16 4GU
Tel: 023 8079 2000
http://www.ordnancesurvey.co.uk

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

Reply | Threaded
Open this post in threaded view
|

Re: reasonning question

Rinke Hoekstra-3
In reply to this post by Iulian Alecu
Hi,

It is probably because CP1 cannot be a subclass of CP3 because A is a
superclass of B.

Regards,

        Rinke

Iulian Alecu wrote:

>  
> Hi,
>
> I have the following situation for which I can't give my self an explanation
> :
>  
> B partial A;
> D partial C;
>
> Restrictions R1 and R2 ;
>
> CP1 complete (R1 some A) or (R2 some C);
>
> CP2 complete (R1 some B) or (R2 some D);
>
> At this point I agree that CP2 inferred_subclassof CP1;
>
> Now, if I define
> CP3 complete (R1 some B) or (R2 some owl:Thing);
>
>  I would expect to get after classification that both CP1 and CP2 are
> subclasses of CP3, bu actually I get just CP2 as subclass of CP3.
>  I feel that I'm missing an important point about reasonning in OWL but I
> just can't figure out which.
>
> Thanks ,
>
> Iulian Alecu
> -----------------------------------
> Laboratoire SPIM
> INSERM U729
> -----------------------------------  
>
> -------------------------------------------------------------------------
> To unsubscribe go to http://protege.stanford.edu/community/subscribe.html
>
>

--
--------------------------------------
Rinke Hoekstra         [hidden email]
T: +31-20-5253499    F: +31-20-5253495
Leibniz Center for Law,    Law Faculty
University of Amsterdam,   PO Box 1030
1000 BA  Amsterdam,    The Netherlands
--------------------------------------

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

Reply | Threaded
Open this post in threaded view
|

Re: reasonning question

Ulrike Sattler
In reply to this post by Iulian Alecu

On 6 Mar 2006, at 14:52, Iulian Alecu wrote:

>
> Hi,
>
> I have the following situation for which I can't give my self an  
> explanation
> :
>
> B partial A;
> D partial C;
>
> Restrictions R1 and R2 ;
>
> CP1 complete (R1 some A) or (R2 some C);
>
> CP2 complete (R1 some B) or (R2 some D);
>
> At this point I agree that CP2 inferred_subclassof CP1;

I agree as well

>
> Now, if I define
> CP3 complete (R1 some B) or (R2 some owl:Thing);
>

>  I would expect to get after classification that both CP1 and CP2 are
> subclasses of CP3, bu actually I get just CP2 as subclass of CP3.

indeed, this is correct

>  I feel that I'm missing an important point about reasonning in OWL  
> but I
> just can't figure out which.

ok, so you say that

- every instance of CP2 has an R1-successor in B or an R2-successor  
in D, and that
- every instance of CP3 has an R1-successor in B or an R2-successor  
that can be anything.

Hence CP2 is a subclass of CP3 (because being an instance of D  
clearly implies being an instance of anything/owl:Thing)

CP3 is NOT a cubclass of CP2 because we can think of some x which is  
an instance of CP3 because it has an R2-successor y which is an  
instance of owl:thing, but not of D. Now, if x has no R1-successor at  
all (which would be fine for it being an instance of CP3), then  x is  
NOT an instance of CP2.

Does this help? Cheers, Uli

>
> Thanks ,
>
> Iulian Alecu
> -----------------------------------
> Laboratoire SPIM
> INSERM U729
> -----------------------------------
>
> ----------------------------------------------------------------------
> ---
> To unsubscribe go to http://protege.stanford.edu/community/ 
> subscribe.html
>

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

Reply | Threaded
Open this post in threaded view
|

Re: reasonning question

Iulian Alecu
Uli,Hai,John

Thanks for your answers and sorry for multiple posting.

Uli's explanation made me understand. The thing that bothered me was that I
have an "or" who's acting as an "and" (well or as I just sad I'm missing a
point).

John in his answer proposing a model having intersections :

Male partial Person
Cat partial Animal

CP1 = hasChild some Person and hasPet some Animal
CP2 = hasChild some Male and hasPet some Cat

CP3 = hasChild some Male and hasPet some owl:Thing

(an more proper examples, thanks John), but actually he used replaced by
"and" the "or" in my example.

Actually, this is the behavior that I expect from a and model not from a
"or".
Could anyone tell me why "or" and "and" looks like having the same result ?

Regards,

iulian


-----Message d'origine-----
De : [hidden email]
[mailto:[hidden email]] De la part de Uli Sattler
Envoyé : lundi 6 mars 2006 16:49
À : [hidden email]
Objet : [protege-owl] Re: reasonning question


On 6 Mar 2006, at 14:52, Iulian Alecu wrote:

>
> Hi,
>
> I have the following situation for which I can't give my self an
> explanation
> :
>
> B partial A;
> D partial C;
>
> Restrictions R1 and R2 ;
>
> CP1 complete (R1 some A) or (R2 some C);
>
> CP2 complete (R1 some B) or (R2 some D);
>
> At this point I agree that CP2 inferred_subclassof CP1;

I agree as well

>
> Now, if I define
> CP3 complete (R1 some B) or (R2 some owl:Thing);
>

>  I would expect to get after classification that both CP1 and CP2 are
> subclasses of CP3, bu actually I get just CP2 as subclass of CP3.

indeed, this is correct

>  I feel that I'm missing an important point about reasonning in OWL
> but I just can't figure out which.

ok, so you say that

- every instance of CP2 has an R1-successor in B or an R2-successor in D,
and that
- every instance of CP3 has an R1-successor in B or an R2-successor that can
be anything.

Hence CP2 is a subclass of CP3 (because being an instance of D clearly
implies being an instance of anything/owl:Thing)

CP3 is NOT a cubclass of CP2 because we can think of some x which is an
instance of CP3 because it has an R2-successor y which is an instance of
owl:thing, but not of D. Now, if x has no R1-successor at all (which would
be fine for it being an instance of CP3), then  x is NOT an instance of CP2.

Does this help? Cheers, Uli

>
> Thanks ,
>
> Iulian Alecu
> -----------------------------------
> Laboratoire SPIM
> INSERM U729
> -----------------------------------
>
> ----------------------------------------------------------------------
> ---
> To unsubscribe go to http://protege.stanford.edu/community/
> subscribe.html
>

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

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

Reply | Threaded
Open this post in threaded view
|

Re: reasonning question

Ulrike Sattler

On 6 Mar 2006, at 16:40, Iulian Alecu wrote:

> Uli,Hai,John
>
> Thanks for your answers and sorry for multiple posting.
>
> Uli's explanation made me understand. The thing that bothered me  
> was that I
> have an "or" who's acting as an "and" (well or as I just sad I'm  
> missing a
> point).
>
> John in his answer proposing a model having intersections :
>
> Male partial Person
> Cat partial Animal
>
> CP1 = hasChild some Person and hasPet some Animal
> CP2 = hasChild some Male and hasPet some Cat
>
> CP3 = hasChild some Male and hasPet some owl:Thing
>
> (an more proper examples, thanks John), but actually he used  
> replaced by
> "and" the "or" in my example.
>
> Actually, this is the behavior that I expect from a and model not  
> from a
> "or".
> Could anyone tell me why "or" and "and" looks like having the same  
> result ?
>

because they dont' matter in this case: what matters is that every  
cat is a thing, but not every thing is a cat.....

Cheers, Uli

> Regards,
>
> iulian
>
>
> -----Message d'origine-----
> De : [hidden email]
> [mailto:[hidden email]] De la part de Uli  
> Sattler
> Envoyé : lundi 6 mars 2006 16:49
> À : [hidden email]
> Objet : [protege-owl] Re: reasonning question
>
>
> On 6 Mar 2006, at 14:52, Iulian Alecu wrote:
>
>>
>> Hi,
>>
>> I have the following situation for which I can't give my self an
>> explanation
>> :
>>
>> B partial A;
>> D partial C;
>>
>> Restrictions R1 and R2 ;
>>
>> CP1 complete (R1 some A) or (R2 some C);
>>
>> CP2 complete (R1 some B) or (R2 some D);
>>
>> At this point I agree that CP2 inferred_subclassof CP1;
>
> I agree as well
>
>>
>> Now, if I define
>> CP3 complete (R1 some B) or (R2 some owl:Thing);
>>
>
>>  I would expect to get after classification that both CP1 and CP2 are
>> subclasses of CP3, bu actually I get just CP2 as subclass of CP3.
>
> indeed, this is correct
>
>>  I feel that I'm missing an important point about reasonning in OWL
>> but I just can't figure out which.
>
> ok, so you say that
>
> - every instance of CP2 has an R1-successor in B or an R2-successor  
> in D,
> and that
> - every instance of CP3 has an R1-successor in B or an R2-successor  
> that can
> be anything.
>
> Hence CP2 is a subclass of CP3 (because being an instance of D clearly
> implies being an instance of anything/owl:Thing)
>
> CP3 is NOT a cubclass of CP2 because we can think of some x which  
> is an
> instance of CP3 because it has an R2-successor y which is an  
> instance of
> owl:thing, but not of D. Now, if x has no R1-successor at all  
> (which would
> be fine for it being an instance of CP3), then  x is NOT an  
> instance of CP2.
>
> Does this help? Cheers, Uli
>
>>
>> Thanks ,
>>
>> Iulian Alecu
>> -----------------------------------
>> Laboratoire SPIM
>> INSERM U729
>> -----------------------------------
>>
>> ---------------------------------------------------------------------
>> -
>> ---
>> To unsubscribe go to http://protege.stanford.edu/community/
>> subscribe.html
>>
>
> ----------------------------------------------------------------------
> ---
> To unsubscribe go to http://protege.stanford.edu/community/ 
> subscribe.html
>
> ----------------------------------------------------------------------
> ---
> To unsubscribe go to http://protege.stanford.edu/community/ 
> subscribe.html
>

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

Reply | Threaded
Open this post in threaded view
|

Re: reasonning question

Raj M Verma
In reply to this post by Ulrike Sattler
hi Uli,

your explanation is very clear... I guess the question of Lulian is a little bit different... You tried to explain why CP3 in not a subclass of CP2, but (I assume) Lulian wants to know why CP1 is not a subclass of CP3 !!

I'll try to explain this with a pictorial representation... Uli, pls correct if I'm wrong...

CP1:
                T
              /    \
       R1 /        \ R2
          /   OR    \
       A               C


CP2:
                T
              /    \
       R1 /        \ R2
          /   OR    \
       B               D


CP3:
                T
              /    \
       R1 /        \ R2
          /   OR    \
       B               T


Now, to show that CP1 is NOT a subclass of CP3, we have to consider 'both' the cases of R1 and R2... For CP1, every instance of it has an R2-successor in C, and for CP3, every instance of it has an R2-successor in T(Owl:Thing)... so here, for R2, "C partial T" holds...

Take the other case... For CP1, every instance of it has an R1-successor in A, and for CP3, every instance of it has an R1-successor in B... but here, for R1, "A partial B" doesn't hold... it's just the opposite of what u've defined...

I think you must have confused with the "OR" between R1 and R2, where one might assume that if either of the cases holds then we can claim that "CP1 partial CP3" holds... it's not correct... every case shud hold, doesn't matter if it has OR or AND between different roles...

hth,
Raj.
Reply | Threaded
Open this post in threaded view
|

Re: reasonning question

Iulian Alecu
In reply to this post by Ulrike Sattler
Uli,


I feel becoming ennoying, but I'm learning a lot ...
But in my example :

B partial A;
D partial C;

CP1 complete (R1 some A) or (R2 some C);

CP2 complete (R1 some B) or (R2 some D);

CP3 complete (R1 some B) or (R2 some owl:Thing);

If I define

CP4 complete (R1 some B) AND (R2 some owl:Thing);

Would I get CP3 equiv CP4 ? (actually I know that I wouldn't but why ?)

Regards
Iulian





-----Message d'origine-----
De : [hidden email]
[mailto:[hidden email]] De la part de Uli Sattler
Envoyé : lundi 6 mars 2006 17:58
À : [hidden email]
Objet : [protege-owl] Re: reasonning question


On 6 Mar 2006, at 16:40, Iulian Alecu wrote:

> Uli,Hai,John
>
> Thanks for your answers and sorry for multiple posting.
>
> Uli's explanation made me understand. The thing that bothered me was
> that I have an "or" who's acting as an "and" (well or as I just sad
> I'm missing a point).
>
> John in his answer proposing a model having intersections :
>
> Male partial Person
> Cat partial Animal
>
> CP1 = hasChild some Person and hasPet some Animal
> CP2 = hasChild some Male and hasPet some Cat
>
> CP3 = hasChild some Male and hasPet some owl:Thing
>
> (an more proper examples, thanks John), but actually he used replaced
> by "and" the "or" in my example.
>
> Actually, this is the behavior that I expect from a and model not from
> a "or".
> Could anyone tell me why "or" and "and" looks like having the same
> result ?
>

because they dont' matter in this case: what matters is that every cat is a
thing, but not every thing is a cat.....

Cheers, Uli

> Regards,
>
> iulian
>
>
> -----Message d'origine-----
> De : [hidden email]
> [mailto:[hidden email]] De la part de Uli
> Sattler Envoyé : lundi 6 mars 2006 16:49 À :
> [hidden email] Objet : [protege-owl] Re: reasonning
> question
>
>
> On 6 Mar 2006, at 14:52, Iulian Alecu wrote:
>
>>
>> Hi,
>>
>> I have the following situation for which I can't give my self an
>> explanation
>> :
>>
>> B partial A;
>> D partial C;
>>
>> Restrictions R1 and R2 ;
>>
>> CP1 complete (R1 some A) or (R2 some C);
>>
>> CP2 complete (R1 some B) or (R2 some D);
>>
>> At this point I agree that CP2 inferred_subclassof CP1;
>
> I agree as well
>
>>
>> Now, if I define
>> CP3 complete (R1 some B) or (R2 some owl:Thing);
>>
>
>>  I would expect to get after classification that both CP1 and CP2 are
>> subclasses of CP3, bu actually I get just CP2 as subclass of CP3.
>
> indeed, this is correct
>
>>  I feel that I'm missing an important point about reasonning in OWL
>> but I just can't figure out which.
>
> ok, so you say that
>
> - every instance of CP2 has an R1-successor in B or an R2-successor in
> D, and that
> - every instance of CP3 has an R1-successor in B or an R2-successor
> that can be anything.
>
> Hence CP2 is a subclass of CP3 (because being an instance of D clearly
> implies being an instance of anything/owl:Thing)
>
> CP3 is NOT a cubclass of CP2 because we can think of some x which is
> an instance of CP3 because it has an R2-successor y which is an
> instance of owl:thing, but not of D. Now, if x has no R1-successor at
> all (which would be fine for it being an instance of CP3), then  x is
> NOT an instance of CP2.
>
> Does this help? Cheers, Uli
>
>>
>> Thanks ,
>>
>> Iulian Alecu
>> -----------------------------------
>> Laboratoire SPIM
>> INSERM U729
>> -----------------------------------
>>
>> ---------------------------------------------------------------------
>> -
>> ---
>> To unsubscribe go to http://protege.stanford.edu/community/
>> subscribe.html
>>
>
> ----------------------------------------------------------------------
> ---
> To unsubscribe go to http://protege.stanford.edu/community/
> subscribe.html
>
> ----------------------------------------------------------------------
> ---
> To unsubscribe go to http://protege.stanford.edu/community/
> subscribe.html
>

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

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

Reply | Threaded
Open this post in threaded view
|

Re: reasonning question

Iulian Alecu
In reply to this post by Raj M Verma
Raj ,
 
Thanks, I got the idea from the uli's explanation, your's make it to me even more clear.
 
But actually I get a different behaviour if I put AND. 
 
if  I  define CP4 as
                T
              /    \
       R1 /        \ R2
          /  AND  \
       B               T

 I dont get CP3 equiv CP4 !
 
regards Iulian
 


De : [hidden email] [mailto:[hidden email]] De la part de Raj M Verma
Envoyé : lundi 6 mars 2006 18:12
À : [hidden email]
Objet : [protege-owl] Re: reasonning question

hi Uli,

your explanation is very clear... I guess the question of Lulian is a little bit different... You tried to explain why CP3 in not a subclass of CP2, but (I assume) Lulian wants to know why CP1 is not a subclass of CP3 !!

I'll try to explain this with a pictorial representation... Uli, pls correct if I'm wrong...

CP1:
                T
              /    \
       R1 /        \ R2
          /   OR    \
       A               C


CP2:
                T
              /    \
       R1 /        \ R2
          /   OR    \
       B               D


CP3:
                T
              /    \
       R1 /        \ R2
          /   OR    \
       B               T


Now, to show that CP1 is NOT a subclass of CP3, we have to consider 'both' the cases of R1 and R2... For CP1, every instance of it has an R2-successor in C, and for CP3, every instance of it has an R2-successor in T(Owl:Thing)... so here, for R2, "C partial T" holds...

Take the other case... For CP1, every instance of it has an R1-successor in A, and for CP3, every instance of it has an R1-successor in B... but here, for R1, "A partial B" doesn't hold... it's just the opposite of what u've defined...

I think you must have confused with the "OR" between R1 and R2, where one might assume that if either of the cases holds then we can claim that "CP1 partial CP3" holds... it's not correct... every case shud hold, doesn't matter if it has OR or AND between different roles...

hth,
Raj.
Reply | Threaded
Open this post in threaded view
|

Re: reasonning question

Ulrike Sattler
In reply to this post by Iulian Alecu

On 6 Mar 2006, at 17:19, Iulian Alecu wrote:

> Uli,
>
>
> I feel becoming ennoying, but I'm learning a lot ...
> But in my example :
>
> B partial A;
> D partial C;
>
> CP1 complete (R1 some A) or (R2 some C);
>
> CP2 complete (R1 some B) or (R2 some D);
>
> CP3 complete (R1 some B) or (R2 some owl:Thing);
>
> If I define
>
> CP4 complete (R1 some B) AND (R2 some owl:Thing);
>
> Would I get CP3 equiv CP4 ? (actually I know that I wouldn't but  
> why ?)
>

hm, we have clearly that CP4 is a subclass of C3, but not vice versa:

I can easily find an individual, say x, which is an instance of CP3,  
but NOT of CP4: x has an R1-successor who is an instance of B, but  
does not have any R2-successors at all. Hence x satisfies one of the  
disjuncts of CP3, but not both of the conjuncts of CP4, and thus it  
is an instance of CP3, but not an instance of CP4.

Cheers, Uli

> Regards
> Iulian
>
>
>
>
>
> -----Message d'origine-----
> De : [hidden email]
> [mailto:[hidden email]] De la part de Uli  
> Sattler
> Envoyé : lundi 6 mars 2006 17:58
> À : [hidden email]
> Objet : [protege-owl] Re: reasonning question
>
>
> On 6 Mar 2006, at 16:40, Iulian Alecu wrote:
>
>> Uli,Hai,John
>>
>> Thanks for your answers and sorry for multiple posting.
>>
>> Uli's explanation made me understand. The thing that bothered me was
>> that I have an "or" who's acting as an "and" (well or as I just sad
>> I'm missing a point).
>>
>> John in his answer proposing a model having intersections :
>>
>> Male partial Person
>> Cat partial Animal
>>
>> CP1 = hasChild some Person and hasPet some Animal
>> CP2 = hasChild some Male and hasPet some Cat
>>
>> CP3 = hasChild some Male and hasPet some owl:Thing
>>
>> (an more proper examples, thanks John), but actually he used replaced
>> by "and" the "or" in my example.
>>
>> Actually, this is the behavior that I expect from a and model not  
>> from
>> a "or".
>> Could anyone tell me why "or" and "and" looks like having the same
>> result ?
>>
>
> because they dont' matter in this case: what matters is that every  
> cat is a
> thing, but not every thing is a cat.....
>
> Cheers, Uli
>
>> Regards,
>>
>> iulian
>>
>>
>> -----Message d'origine-----
>> De : [hidden email]
>> [mailto:[hidden email]] De la part de Uli
>> Sattler Envoyé : lundi 6 mars 2006 16:49 À :
>> [hidden email] Objet : [protege-owl] Re: reasonning
>> question
>>
>>
>> On 6 Mar 2006, at 14:52, Iulian Alecu wrote:
>>
>>>
>>> Hi,
>>>
>>> I have the following situation for which I can't give my self an
>>> explanation
>>> :
>>>
>>> B partial A;
>>> D partial C;
>>>
>>> Restrictions R1 and R2 ;
>>>
>>> CP1 complete (R1 some A) or (R2 some C);
>>>
>>> CP2 complete (R1 some B) or (R2 some D);
>>>
>>> At this point I agree that CP2 inferred_subclassof CP1;
>>
>> I agree as well
>>
>>>
>>> Now, if I define
>>> CP3 complete (R1 some B) or (R2 some owl:Thing);
>>>
>>
>>>  I would expect to get after classification that both CP1 and CP2  
>>> are
>>> subclasses of CP3, bu actually I get just CP2 as subclass of CP3.
>>
>> indeed, this is correct
>>
>>>  I feel that I'm missing an important point about reasonning in OWL
>>> but I just can't figure out which.
>>
>> ok, so you say that
>>
>> - every instance of CP2 has an R1-successor in B or an R2-
>> successor in
>> D, and that
>> - every instance of CP3 has an R1-successor in B or an R2-successor
>> that can be anything.
>>
>> Hence CP2 is a subclass of CP3 (because being an instance of D  
>> clearly
>> implies being an instance of anything/owl:Thing)
>>
>> CP3 is NOT a cubclass of CP2 because we can think of some x which is
>> an instance of CP3 because it has an R2-successor y which is an
>> instance of owl:thing, but not of D. Now, if x has no R1-successor at
>> all (which would be fine for it being an instance of CP3), then  x is
>> NOT an instance of CP2.
>>
>> Does this help? Cheers, Uli
>>
>>>
>>> Thanks ,
>>>
>>> Iulian Alecu
>>> -----------------------------------
>>> Laboratoire SPIM
>>> INSERM U729
>>> -----------------------------------
>>>
>>> --------------------------------------------------------------------
>>> -
>>> -
>>> ---
>>> To unsubscribe go to http://protege.stanford.edu/community/
>>> subscribe.html
>>>
>>
>> ---------------------------------------------------------------------
>> -
>> ---
>> To unsubscribe go to http://protege.stanford.edu/community/
>> subscribe.html
>>
>> ---------------------------------------------------------------------
>> -
>> ---
>> To unsubscribe go to http://protege.stanford.edu/community/
>> subscribe.html
>>
>
> ----------------------------------------------------------------------
> ---
> To unsubscribe go to http://protege.stanford.edu/community/ 
> subscribe.html
>
> ----------------------------------------------------------------------
> ---
> To unsubscribe go to http://protege.stanford.edu/community/ 
> subscribe.html
>

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

Reply | Threaded
Open this post in threaded view
|

Re: reasonning question

Raj M Verma
In reply to this post by Iulian Alecu
hi Lulian,

CP3:
                T
              /    \
       R1 /        \ R2
          /   OR    \
       B               T

CP4:
                T
              /    \
       R1 /        \ R2
          /  AND   \
       B               T

CP3 and CP4 are not equivalent b'cos, for every instance of CP4 there must be R1 and R2, but for CP3 there can be a case where only R1 is used but not R2...

hth,
Raj.
Reply | Threaded
Open this post in threaded view
|

Re: reasonning question

Michael Latta
In reply to this post by Iulian Alecu
CP3 is not equiv CP4, and CP3 is not a superset of CP3.  Think set theory
not OO class inheritance.  The set of objects allowed in CP3 is larger than
the set of objects in CP1 so CP1 can not be a sub set of CP3.

Michael

-----Original Message-----
From: [hidden email]
[mailto:[hidden email]] On Behalf Of Iulian Alecu
Sent: Monday, March 06, 2006 9:19 AM
To: [hidden email]
Subject: [protege-owl] Re: reasonning question

Uli,


I feel becoming ennoying, but I'm learning a lot ...
But in my example :

B partial A;
D partial C;

CP1 complete (R1 some A) or (R2 some C);

CP2 complete (R1 some B) or (R2 some D);

CP3 complete (R1 some B) or (R2 some owl:Thing);

If I define

CP4 complete (R1 some B) AND (R2 some owl:Thing);

Would I get CP3 equiv CP4 ? (actually I know that I wouldn't but why ?)

Regards
Iulian





-----Message d'origine-----
De : [hidden email]
[mailto:[hidden email]] De la part de Uli Sattler
Envoyé : lundi 6 mars 2006 17:58
À : [hidden email]
Objet : [protege-owl] Re: reasonning question


On 6 Mar 2006, at 16:40, Iulian Alecu wrote:

> Uli,Hai,John
>
> Thanks for your answers and sorry for multiple posting.
>
> Uli's explanation made me understand. The thing that bothered me was
> that I have an "or" who's acting as an "and" (well or as I just sad
> I'm missing a point).
>
> John in his answer proposing a model having intersections :
>
> Male partial Person
> Cat partial Animal
>
> CP1 = hasChild some Person and hasPet some Animal
> CP2 = hasChild some Male and hasPet some Cat
>
> CP3 = hasChild some Male and hasPet some owl:Thing
>
> (an more proper examples, thanks John), but actually he used replaced
> by "and" the "or" in my example.
>
> Actually, this is the behavior that I expect from a and model not from
> a "or".
> Could anyone tell me why "or" and "and" looks like having the same
> result ?
>

because they dont' matter in this case: what matters is that every cat is a
thing, but not every thing is a cat.....

Cheers, Uli

> Regards,
>
> iulian
>
>
> -----Message d'origine-----
> De : [hidden email]
> [mailto:[hidden email]] De la part de Uli
> Sattler Envoyé : lundi 6 mars 2006 16:49 À :
> [hidden email] Objet : [protege-owl] Re: reasonning
> question
>
>
> On 6 Mar 2006, at 14:52, Iulian Alecu wrote:
>
>>
>> Hi,
>>
>> I have the following situation for which I can't give my self an
>> explanation
>> :
>>
>> B partial A;
>> D partial C;
>>
>> Restrictions R1 and R2 ;
>>
>> CP1 complete (R1 some A) or (R2 some C);
>>
>> CP2 complete (R1 some B) or (R2 some D);
>>
>> At this point I agree that CP2 inferred_subclassof CP1;
>
> I agree as well
>
>>
>> Now, if I define
>> CP3 complete (R1 some B) or (R2 some owl:Thing);
>>
>
>>  I would expect to get after classification that both CP1 and CP2 are
>> subclasses of CP3, bu actually I get just CP2 as subclass of CP3.
>
> indeed, this is correct
>
>>  I feel that I'm missing an important point about reasonning in OWL
>> but I just can't figure out which.
>
> ok, so you say that
>
> - every instance of CP2 has an R1-successor in B or an R2-successor in
> D, and that
> - every instance of CP3 has an R1-successor in B or an R2-successor
> that can be anything.
>
> Hence CP2 is a subclass of CP3 (because being an instance of D clearly
> implies being an instance of anything/owl:Thing)
>
> CP3 is NOT a cubclass of CP2 because we can think of some x which is
> an instance of CP3 because it has an R2-successor y which is an
> instance of owl:thing, but not of D. Now, if x has no R1-successor at
> all (which would be fine for it being an instance of CP3), then  x is
> NOT an instance of CP2.
>
> Does this help? Cheers, Uli
>
>>
>> Thanks ,
>>
>> Iulian Alecu
>> -----------------------------------
>> Laboratoire SPIM
>> INSERM U729
>> -----------------------------------
>>
>> ---------------------------------------------------------------------
>> -
>> ---
>> To unsubscribe go to http://protege.stanford.edu/community/
>> subscribe.html
>>
>
> ----------------------------------------------------------------------
> ---
> To unsubscribe go to http://protege.stanford.edu/community/
> subscribe.html
>
> ----------------------------------------------------------------------
> ---
> To unsubscribe go to http://protege.stanford.edu/community/
> subscribe.html
>

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

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

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

Reply | Threaded
Open this post in threaded view
|

Re: reasonning question

John Goodwin
In reply to this post by Iulian Alecu
Sorry for the confusion iulian that was a typo - on my behalf, meant to right "or" not "and". Glad Uli sorted the confusion out for you.
 
John

________________________________

From: [hidden email] on behalf of Iulian Alecu
Sent: Mon 06/03/2006 16:40
To: [hidden email]
Subject: [protege-owl] Re: reasonning question



Uli,Hai,John

Thanks for your answers and sorry for multiple posting.

Uli's explanation made me understand. The thing that bothered me was that I
have an "or" who's acting as an "and" (well or as I just sad I'm missing a
point).

John in his answer proposing a model having intersections :

Male partial Person
Cat partial Animal

CP1 = hasChild some Person and hasPet some Animal
CP2 = hasChild some Male and hasPet some Cat

CP3 = hasChild some Male and hasPet some owl:Thing

(an more proper examples, thanks John), but actually he used replaced by
"and" the "or" in my example.

Actually, this is the behavior that I expect from a and model not from a
"or".
Could anyone tell me why "or" and "and" looks like having the same result ?

Regards,

iulian


-----Message d'origine-----
De : [hidden email]
[mailto:[hidden email]] De la part de Uli Sattler
Envoyé : lundi 6 mars 2006 16:49
À : [hidden email]
Objet : [protege-owl] Re: reasonning question


On 6 Mar 2006, at 14:52, Iulian Alecu wrote:

>
> Hi,
>
> I have the following situation for which I can't give my self an
> explanation
> :
>
> B partial A;
> D partial C;
>
> Restrictions R1 and R2 ;
>
> CP1 complete (R1 some A) or (R2 some C);
>
> CP2 complete (R1 some B) or (R2 some D);
>
> At this point I agree that CP2 inferred_subclassof CP1;
I agree as well

>
> Now, if I define
>       CP3 complete (R1 some B) or (R2 some owl:Thing);
>

>  I would expect to get after classification that both CP1 and CP2 are
> subclasses of CP3, bu actually I get just CP2 as subclass of CP3.

indeed, this is correct

>  I feel that I'm missing an important point about reasonning in OWL
> but I just can't figure out which.

ok, so you say that

- every instance of CP2 has an R1-successor in B or an R2-successor in D,
and that
- every instance of CP3 has an R1-successor in B or an R2-successor that can
be anything.

Hence CP2 is a subclass of CP3 (because being an instance of D clearly
implies being an instance of anything/owl:Thing)

CP3 is NOT a cubclass of CP2 because we can think of some x which is an
instance of CP3 because it has an R2-successor y which is an instance of
owl:thing, but not of D. Now, if x has no R1-successor at all (which would
be fine for it being an instance of CP3), then  x is NOT an instance of CP2.

Does this help? Cheers, Uli

>
> Thanks ,
>
> Iulian Alecu
> -----------------------------------
> Laboratoire SPIM
> INSERM U729
> -----------------------------------
>
> ----------------------------------------------------------------------
> ---
> To unsubscribe go to http://protege.stanford.edu/community/
> subscribe.html
>
-------------------------------------------------------------------------
To unsubscribe go to http://protege.stanford.edu/community/subscribe.html

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




.


This email is only intended for the person to whom it is addressed and may contain confidential information. If you have received this email in error, please notify the sender and delete this email which must not be copied, distributed or disclosed to any other person.

Unless stated otherwise, the contents of this email are personal to the writer and do not represent the official view of Ordnance Survey. Nor can any contract be formed on Ordnance Survey's behalf via email. We reserve the right to monitor emails and attachments without prior notice.

Thank you for your cooperation.

Ordnance Survey
Romsey Road
Southampton SO16 4GU
Tel: 023 8079 2000
http://www.ordnancesurvey.co.uk

winmail.dat (8K) Download Attachment
Reply | Threaded
Open this post in threaded view
|

automatic ontology generation

Knarig Arabshian
Hello,

Do you know if anyone has worked on automatic ontology generation using protege-owl? For instance, if a protege project has a number of instances, is there a program that looks through these instances and depending on the various property values, automatically creates restricted subclasses where these instances can be classified more specifically? If not, do you know if others have looked at this type of ontology generation? The ones I have found are mostly ontology generation for web content.

Thanks,
Knarig


Yahoo! Mail
Bring photos to life! New PhotoMail makes sharing a breeze.
Reply | Threaded
Open this post in threaded view
|

Re: reasonning question

Iulian Alecu
In reply to this post by Michael Latta
Michael,

Actually, I tried to explain this by the set theory, but i din't now
wethever the subset of objects allowed  to a class by a restriction R1 is
disjoint with the subset allowed by a R2 restriction.

Correct me if I'm wrong, now from all these explanation I concluded that
they are allways disjoint.

Regards
Iulian
 

-----Message d'origine-----
De : [hidden email]
[mailto:[hidden email]] De la part de Michael Latta
Envoyé : lundi 6 mars 2006 18:36
À : [hidden email]
Objet : [protege-owl] Re: reasonning question

CP3 is not equiv CP4, and CP3 is not a superset of CP3.  Think set theory
not OO class inheritance.  The set of objects allowed in CP3 is larger than
the set of objects in CP1 so CP1 can not be a sub set of CP3.

Michael

-----Original Message-----
From: [hidden email]
[mailto:[hidden email]] On Behalf Of Iulian Alecu
Sent: Monday, March 06, 2006 9:19 AM
To: [hidden email]
Subject: [protege-owl] Re: reasonning question

Uli,


I feel becoming ennoying, but I'm learning a lot ...
But in my example :

B partial A;
D partial C;

CP1 complete (R1 some A) or (R2 some C);

CP2 complete (R1 some B) or (R2 some D);

CP3 complete (R1 some B) or (R2 some owl:Thing);

If I define

CP4 complete (R1 some B) AND (R2 some owl:Thing);

Would I get CP3 equiv CP4 ? (actually I know that I wouldn't but why ?)

Regards
Iulian





-----Message d'origine-----
De : [hidden email]
[mailto:[hidden email]] De la part de Uli Sattler
Envoyé : lundi 6 mars 2006 17:58 À : [hidden email] Objet :
[protege-owl] Re: reasonning question


On 6 Mar 2006, at 16:40, Iulian Alecu wrote:

> Uli,Hai,John
>
> Thanks for your answers and sorry for multiple posting.
>
> Uli's explanation made me understand. The thing that bothered me was
> that I have an "or" who's acting as an "and" (well or as I just sad
> I'm missing a point).
>
> John in his answer proposing a model having intersections :
>
> Male partial Person
> Cat partial Animal
>
> CP1 = hasChild some Person and hasPet some Animal
> CP2 = hasChild some Male and hasPet some Cat
>
> CP3 = hasChild some Male and hasPet some owl:Thing
>
> (an more proper examples, thanks John), but actually he used replaced
> by "and" the "or" in my example.
>
> Actually, this is the behavior that I expect from a and model not from
> a "or".
> Could anyone tell me why "or" and "and" looks like having the same
> result ?
>

because they dont' matter in this case: what matters is that every cat is a
thing, but not every thing is a cat.....

Cheers, Uli

> Regards,
>
> iulian
>
>
> -----Message d'origine-----
> De : [hidden email]
> [mailto:[hidden email]] De la part de Uli
> Sattler Envoyé : lundi 6 mars 2006 16:49 À :
> [hidden email] Objet : [protege-owl] Re: reasonning
> question
>
>
> On 6 Mar 2006, at 14:52, Iulian Alecu wrote:
>
>>
>> Hi,
>>
>> I have the following situation for which I can't give my self an
>> explanation
>> :
>>
>> B partial A;
>> D partial C;
>>
>> Restrictions R1 and R2 ;
>>
>> CP1 complete (R1 some A) or (R2 some C);
>>
>> CP2 complete (R1 some B) or (R2 some D);
>>
>> At this point I agree that CP2 inferred_subclassof CP1;
>
> I agree as well
>
>>
>> Now, if I define
>> CP3 complete (R1 some B) or (R2 some owl:Thing);
>>
>
>>  I would expect to get after classification that both CP1 and CP2 are
>> subclasses of CP3, bu actually I get just CP2 as subclass of CP3.
>
> indeed, this is correct
>
>>  I feel that I'm missing an important point about reasonning in OWL
>> but I just can't figure out which.
>
> ok, so you say that
>
> - every instance of CP2 has an R1-successor in B or an R2-successor in
> D, and that
> - every instance of CP3 has an R1-successor in B or an R2-successor
> that can be anything.
>
> Hence CP2 is a subclass of CP3 (because being an instance of D clearly
> implies being an instance of anything/owl:Thing)
>
> CP3 is NOT a cubclass of CP2 because we can think of some x which is
> an instance of CP3 because it has an R2-successor y which is an
> instance of owl:thing, but not of D. Now, if x has no R1-successor at
> all (which would be fine for it being an instance of CP3), then  x is
> NOT an instance of CP2.
>
> Does this help? Cheers, Uli
>
>>
>> Thanks ,
>>
>> Iulian Alecu
>> -----------------------------------
>> Laboratoire SPIM
>> INSERM U729
>> -----------------------------------
>>
>> ---------------------------------------------------------------------
>> -
>> ---
>> To unsubscribe go to http://protege.stanford.edu/community/
>> subscribe.html
>>
>
> ----------------------------------------------------------------------
> ---
> To unsubscribe go to http://protege.stanford.edu/community/
> subscribe.html
>
> ----------------------------------------------------------------------
> ---
> To unsubscribe go to http://protege.stanford.edu/community/
> subscribe.html
>

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

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

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

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

Reply | Threaded
Open this post in threaded view
|

Re: automatic ontology generation

Ulrike Sattler
In reply to this post by Knarig Arabshian
Hi, you might want to have a look at Sonic at 


I don't know how much work it would be 
to make Sonic run with Protege-Owl, but 
Sonic implements very helpful, well-founded
services for (what we call) the bottom-up design
of ontologies. 

Cheers, Uli

On 6 Mar 2006, at 22:15, Knarig Arabshian wrote:

Hello,

Do you know if anyone has worked on automatic ontology generation using protege-owl? For instance, if a protege project has a number of instances, is there a program that looks through these instances and depending on the various property values, automatically creates restricted subclasses where these instances can be classified more specifically? If not, do you know if others have looked at this type of ontology generation? The ones I have found are mostly ontology generation for web content.

Thanks,
Knarig


Yahoo! Mail
Bring photos to life! New PhotoMail makes sharing a breeze.

Reply | Threaded
Open this post in threaded view
|

Re: reasonning question

Ulrike Sattler
In reply to this post by Iulian Alecu
On 7 Mar 2006, at 12:00, Iulian Alecu wrote:

> Michael,
>
> Actually, I tried to explain this by the set theory, but i din't now
> wethever the subset of objects allowed  to a class by a restriction  
> R1 is
> disjoint with the subset allowed by a R2 restriction.
>
> Correct me if I'm wrong, now from all these explanation I concluded  
> that
> they are allways disjoint.

wrong -- they may be stated as being disjoint but, in general,
they may overlap. E.g., it is possible to say

CP6 complete (R some A)
CP7 complete (R some A)

Then you can have an instance y of A that is an R-successor of some  
instance x1 of CP6 and that is, at the same time, an R-successor of  
some instance x2 of CP7. But you can also have some instances of A  
that are not R-successors of anybody, neither from CP6 or CP7...or  
just of 1, but not of the other....

Cheers, Uli

>
> Regards
> Iulian
>
>
> -----Message d'origine-----
> De : [hidden email]
> [mailto:[hidden email]] De la part de  
> Michael Latta
> Envoyé : lundi 6 mars 2006 18:36
> À : [hidden email]
> Objet : [protege-owl] Re: reasonning question
>
> CP3 is not equiv CP4, and CP3 is not a superset of CP3.  Think set  
> theory
> not OO class inheritance.  The set of objects allowed in CP3 is  
> larger than
> the set of objects in CP1 so CP1 can not be a sub set of CP3.
>
> Michael
>
> -----Original Message-----
> From: [hidden email]
> [mailto:[hidden email]] On Behalf Of Iulian  
> Alecu
> Sent: Monday, March 06, 2006 9:19 AM
> To: [hidden email]
> Subject: [protege-owl] Re: reasonning question
>
> Uli,
>
>
> I feel becoming ennoying, but I'm learning a lot ...
> But in my example :
>
> B partial A;
> D partial C;
>
> CP1 complete (R1 some A) or (R2 some C);
>
> CP2 complete (R1 some B) or (R2 some D);
>
> CP3 complete (R1 some B) or (R2 some owl:Thing);
>
> If I define
>
> CP4 complete (R1 some B) AND (R2 some owl:Thing);
>
> Would I get CP3 equiv CP4 ? (actually I know that I wouldn't but  
> why ?)
>
> Regards
> Iulian
>
>
>
>
>
> -----Message d'origine-----
> De : [hidden email]
> [mailto:[hidden email]] De la part de Uli  
> Sattler
> Envoyé : lundi 6 mars 2006 17:58 À : [hidden email]  
> Objet :
> [protege-owl] Re: reasonning question
>
>
> On 6 Mar 2006, at 16:40, Iulian Alecu wrote:
>
>> Uli,Hai,John
>>
>> Thanks for your answers and sorry for multiple posting.
>>
>> Uli's explanation made me understand. The thing that bothered me was
>> that I have an "or" who's acting as an "and" (well or as I just sad
>> I'm missing a point).
>>
>> John in his answer proposing a model having intersections :
>>
>> Male partial Person
>> Cat partial Animal
>>
>> CP1 = hasChild some Person and hasPet some Animal
>> CP2 = hasChild some Male and hasPet some Cat
>>
>> CP3 = hasChild some Male and hasPet some owl:Thing
>>
>> (an more proper examples, thanks John), but actually he used replaced
>> by "and" the "or" in my example.
>>
>> Actually, this is the behavior that I expect from a and model not  
>> from
>> a "or".
>> Could anyone tell me why "or" and "and" looks like having the same
>> result ?
>>
>
> because they dont' matter in this case: what matters is that every  
> cat is a
> thing, but not every thing is a cat.....
>
> Cheers, Uli
>
>> Regards,
>>
>> iulian
>>
>>
>> -----Message d'origine-----
>> De : [hidden email]
>> [mailto:[hidden email]] De la part de Uli
>> Sattler Envoyé : lundi 6 mars 2006 16:49 À :
>> [hidden email] Objet : [protege-owl] Re: reasonning
>> question
>>
>>
>> On 6 Mar 2006, at 14:52, Iulian Alecu wrote:
>>
>>>
>>> Hi,
>>>
>>> I have the following situation for which I can't give my self an
>>> explanation
>>> :
>>>
>>> B partial A;
>>> D partial C;
>>>
>>> Restrictions R1 and R2 ;
>>>
>>> CP1 complete (R1 some A) or (R2 some C);
>>>
>>> CP2 complete (R1 some B) or (R2 some D);
>>>
>>> At this point I agree that CP2 inferred_subclassof CP1;
>>
>> I agree as well
>>
>>>
>>> Now, if I define
>>> CP3 complete (R1 some B) or (R2 some owl:Thing);
>>>
>>
>>>  I would expect to get after classification that both CP1 and CP2  
>>> are
>>> subclasses of CP3, bu actually I get just CP2 as subclass of CP3.
>>
>> indeed, this is correct
>>
>>>  I feel that I'm missing an important point about reasonning in OWL
>>> but I just can't figure out which.
>>
>> ok, so you say that
>>
>> - every instance of CP2 has an R1-successor in B or an R2-
>> successor in
>> D, and that
>> - every instance of CP3 has an R1-successor in B or an R2-successor
>> that can be anything.
>>
>> Hence CP2 is a subclass of CP3 (because being an instance of D  
>> clearly
>> implies being an instance of anything/owl:Thing)
>>
>> CP3 is NOT a cubclass of CP2 because we can think of some x which is
>> an instance of CP3 because it has an R2-successor y which is an
>> instance of owl:thing, but not of D. Now, if x has no R1-successor at
>> all (which would be fine for it being an instance of CP3), then  x is
>> NOT an instance of CP2.
>>
>> Does this help? Cheers, Uli
>>
>>>
>>> Thanks ,
>>>
>>> Iulian Alecu
>>> -----------------------------------
>>> Laboratoire SPIM
>>> INSERM U729
>>> -----------------------------------
>>>
>>> --------------------------------------------------------------------
>>> -
>>> -
>>> ---
>>> To unsubscribe go to http://protege.stanford.edu/community/
>>> subscribe.html
>>>
>>
>> ---------------------------------------------------------------------
>> -
>> ---
>> To unsubscribe go to http://protege.stanford.edu/community/
>> subscribe.html
>>
>> ---------------------------------------------------------------------
>> -
>> ---
>> To unsubscribe go to http://protege.stanford.edu/community/
>> subscribe.html
>>
>
> ----------------------------------------------------------------------
> ---
> To unsubscribe go to http://protege.stanford.edu/community/ 
> subscribe.html
>
> ----------------------------------------------------------------------
> ---
> To unsubscribe go to http://protege.stanford.edu/community/ 
> subscribe.html
>
> ----------------------------------------------------------------------
> ---
> To unsubscribe go to http://protege.stanford.edu/community/ 
> subscribe.html
>
> ----------------------------------------------------------------------
> ---
> To unsubscribe go to http://protege.stanford.edu/community/ 
> subscribe.html
>

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

Reply | Threaded
Open this post in threaded view
|

Re: reasonning question

Iulian Alecu
Uli,

I'm ok with your explanation , but actually in the case:

CP8 complete (R1 some A)
CP9 complete (R2 some A)

Taking the risk to be ridiculous, I was wondering if R1 succesors are
dijoint with R2 succesors .
Regards
Iulian

-----Message d'origine-----
De : [hidden email]
[mailto:[hidden email]] De la part de Ulrike Sattler
Envoyé : mardi 7 mars 2006 17:18
À : [hidden email]
Objet : [protege-owl] Re: reasonning question

On 7 Mar 2006, at 12:00, Iulian Alecu wrote:

> Michael,
>
> Actually, I tried to explain this by the set theory, but i din't now
> wethever the subset of objects allowed  to a class by a restriction
> R1 is
> disjoint with the subset allowed by a R2 restriction.
>
> Correct me if I'm wrong, now from all these explanation I concluded
> that they are allways disjoint.

wrong -- they may be stated as being disjoint but, in general, they may
overlap. E.g., it is possible to say

CP6 complete (R some A)
CP7 complete (R some A)

Then you can have an instance y of A that is an R-successor of some instance
x1 of CP6 and that is, at the same time, an R-successor of some instance x2
of CP7. But you can also have some instances of A that are not R-successors
of anybody, neither from CP6 or CP7...or just of 1, but not of the other....

Cheers, Uli

>
> Regards
> Iulian
>
>
> -----Message d'origine-----
> De : [hidden email]
> [mailto:[hidden email]] De la part de Michael
> Latta Envoyé : lundi 6 mars 2006 18:36 À :
> [hidden email] Objet : [protege-owl] Re: reasonning
> question
>
> CP3 is not equiv CP4, and CP3 is not a superset of CP3.  Think set
> theory not OO class inheritance.  The set of objects allowed in CP3 is
> larger than the set of objects in CP1 so CP1 can not be a sub set of
> CP3.
>
> Michael
>
> -----Original Message-----
> From: [hidden email]
> [mailto:[hidden email]] On Behalf Of Iulian
> Alecu
> Sent: Monday, March 06, 2006 9:19 AM
> To: [hidden email]
> Subject: [protege-owl] Re: reasonning question
>
> Uli,
>
>
> I feel becoming ennoying, but I'm learning a lot ...
> But in my example :
>
> B partial A;
> D partial C;
>
> CP1 complete (R1 some A) or (R2 some C);
>
> CP2 complete (R1 some B) or (R2 some D);
>
> CP3 complete (R1 some B) or (R2 some owl:Thing);
>
> If I define
>
> CP4 complete (R1 some B) AND (R2 some owl:Thing);
>
> Would I get CP3 equiv CP4 ? (actually I know that I wouldn't but why
> ?)
>
> Regards
> Iulian
>
>
>
>
>
> -----Message d'origine-----
> De : [hidden email]
> [mailto:[hidden email]] De la part de Uli
> Sattler Envoyé : lundi 6 mars 2006 17:58 À :
> [hidden email] Objet :
> [protege-owl] Re: reasonning question
>
>
> On 6 Mar 2006, at 16:40, Iulian Alecu wrote:
>
>> Uli,Hai,John
>>
>> Thanks for your answers and sorry for multiple posting.
>>
>> Uli's explanation made me understand. The thing that bothered me was
>> that I have an "or" who's acting as an "and" (well or as I just sad
>> I'm missing a point).
>>
>> John in his answer proposing a model having intersections :
>>
>> Male partial Person
>> Cat partial Animal
>>
>> CP1 = hasChild some Person and hasPet some Animal
>> CP2 = hasChild some Male and hasPet some Cat
>>
>> CP3 = hasChild some Male and hasPet some owl:Thing
>>
>> (an more proper examples, thanks John), but actually he used replaced
>> by "and" the "or" in my example.
>>
>> Actually, this is the behavior that I expect from a and model not
>> from a "or".
>> Could anyone tell me why "or" and "and" looks like having the same
>> result ?
>>
>
> because they dont' matter in this case: what matters is that every cat
> is a thing, but not every thing is a cat.....
>
> Cheers, Uli
>
>> Regards,
>>
>> iulian
>>
>>
>> -----Message d'origine-----
>> De : [hidden email]
>> [mailto:[hidden email]] De la part de Uli
>> Sattler Envoyé : lundi 6 mars 2006 16:49 À :
>> [hidden email] Objet : [protege-owl] Re: reasonning
>> question
>>
>>
>> On 6 Mar 2006, at 14:52, Iulian Alecu wrote:
>>
>>>
>>> Hi,
>>>
>>> I have the following situation for which I can't give my self an
>>> explanation
>>> :
>>>
>>> B partial A;
>>> D partial C;
>>>
>>> Restrictions R1 and R2 ;
>>>
>>> CP1 complete (R1 some A) or (R2 some C);
>>>
>>> CP2 complete (R1 some B) or (R2 some D);
>>>
>>> At this point I agree that CP2 inferred_subclassof CP1;
>>
>> I agree as well
>>
>>>
>>> Now, if I define
>>> CP3 complete (R1 some B) or (R2 some owl:Thing);
>>>
>>
>>>  I would expect to get after classification that both CP1 and CP2
>>> are subclasses of CP3, bu actually I get just CP2 as subclass of
>>> CP3.
>>
>> indeed, this is correct
>>
>>>  I feel that I'm missing an important point about reasonning in OWL
>>> but I just can't figure out which.
>>
>> ok, so you say that
>>
>> - every instance of CP2 has an R1-successor in B or an R2- successor
>> in D, and that
>> - every instance of CP3 has an R1-successor in B or an R2-successor
>> that can be anything.
>>
>> Hence CP2 is a subclass of CP3 (because being an instance of D
>> clearly implies being an instance of anything/owl:Thing)
>>
>> CP3 is NOT a cubclass of CP2 because we can think of some x which is
>> an instance of CP3 because it has an R2-successor y which is an
>> instance of owl:thing, but not of D. Now, if x has no R1-successor at
>> all (which would be fine for it being an instance of CP3), then  x is
>> NOT an instance of CP2.
>>
>> Does this help? Cheers, Uli
>>
>>>
>>> Thanks ,
>>>
>>> Iulian Alecu
>>> -----------------------------------
>>> Laboratoire SPIM
>>> INSERM U729
>>> -----------------------------------
>>>
>>> --------------------------------------------------------------------
>>> -
>>> -
>>> ---
>>> To unsubscribe go to http://protege.stanford.edu/community/
>>> subscribe.html
>>>
>>
>> ---------------------------------------------------------------------
>> -
>> ---
>> To unsubscribe go to http://protege.stanford.edu/community/
>> subscribe.html
>>
>> ---------------------------------------------------------------------
>> -
>> ---
>> To unsubscribe go to http://protege.stanford.edu/community/
>> subscribe.html
>>
>
> ----------------------------------------------------------------------
> ---
> To unsubscribe go to http://protege.stanford.edu/community/
> subscribe.html
>
> ----------------------------------------------------------------------
> ---
> To unsubscribe go to http://protege.stanford.edu/community/
> subscribe.html
>
> ----------------------------------------------------------------------
> ---
> To unsubscribe go to http://protege.stanford.edu/community/
> subscribe.html
>
> ----------------------------------------------------------------------
> ---
> To unsubscribe go to http://protege.stanford.edu/community/
> subscribe.html
>

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

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

Reply | Threaded
Open this post in threaded view
|

Re: reasonning question

Ulrike Sattler

On 7 Mar 2006, at 16:59, Iulian Alecu wrote:

> Uli,
>
> I'm ok with your explanation , but actually in the case:
>
> CP8 complete (R1 some A)
> CP9 complete (R2 some A)
>
> Taking the risk to be ridiculous, I was wondering if R1 succesors are
> dijoint with R2 succesors .

No, as before (sorry, I meant to do it like this):

Then you can have an instance y of A that is an R1-successor of some  
instance
x1 of CP8 and that is, at the same time, an R2-successor of some  
instance x2
of CP9. But you can also have some instances of A that are not R1- or  
R2-successors
of anybody, neither from CP8 or CP9...or just of 1, but not of the  
other....

Cheers, Uli

> Regards
> Iulian
>
> -----Message d'origine-----
> De : [hidden email]
> [mailto:[hidden email]] De la part de  
> Ulrike Sattler
> Envoyé : mardi 7 mars 2006 17:18
> À : [hidden email]
> Objet : [protege-owl] Re: reasonning question
>
> On 7 Mar 2006, at 12:00, Iulian Alecu wrote:
>
>> Michael,
>>
>> Actually, I tried to explain this by the set theory, but i din't now
>> wethever the subset of objects allowed  to a class by a restriction
>> R1 is
>> disjoint with the subset allowed by a R2 restriction.
>>
>> Correct me if I'm wrong, now from all these explanation I concluded
>> that they are allways disjoint.
>
> wrong -- they may be stated as being disjoint but, in general, they  
> may
> overlap. E.g., it is possible to say
>
> CP6 complete (R some A)
> CP7 complete (R some A)
>
> Then you can have an instance y of A that is an R-successor of some  
> instance
> x1 of CP6 and that is, at the same time, an R-successor of some  
> instance x2
> of CP7. But you can also have some instances of A that are not R-
> successors
> of anybody, neither from CP6 or CP7...or just of 1, but not of the  
> other....
>
> Cheers, Uli
>
>>
>> Regards
>> Iulian
>>
>>
>> -----Message d'origine-----
>> De : [hidden email]
>> [mailto:[hidden email]] De la part de Michael
>> Latta Envoyé : lundi 6 mars 2006 18:36 À :
>> [hidden email] Objet : [protege-owl] Re: reasonning
>> question
>>
>> CP3 is not equiv CP4, and CP3 is not a superset of CP3.  Think set
>> theory not OO class inheritance.  The set of objects allowed in  
>> CP3 is
>> larger than the set of objects in CP1 so CP1 can not be a sub set of
>> CP3.
>>
>> Michael
>>
>> -----Original Message-----
>> From: [hidden email]
>> [mailto:[hidden email]] On Behalf Of Iulian
>> Alecu
>> Sent: Monday, March 06, 2006 9:19 AM
>> To: [hidden email]
>> Subject: [protege-owl] Re: reasonning question
>>
>> Uli,
>>
>>
>> I feel becoming ennoying, but I'm learning a lot ...
>> But in my example :
>>
>> B partial A;
>> D partial C;
>>
>> CP1 complete (R1 some A) or (R2 some C);
>>
>> CP2 complete (R1 some B) or (R2 some D);
>>
>> CP3 complete (R1 some B) or (R2 some owl:Thing);
>>
>> If I define
>>
>> CP4 complete (R1 some B) AND (R2 some owl:Thing);
>>
>> Would I get CP3 equiv CP4 ? (actually I know that I wouldn't but why
>> ?)
>>
>> Regards
>> Iulian
>>
>>
>>
>>
>>
>> -----Message d'origine-----
>> De : [hidden email]
>> [mailto:[hidden email]] De la part de Uli
>> Sattler Envoyé : lundi 6 mars 2006 17:58 À :
>> [hidden email] Objet :
>> [protege-owl] Re: reasonning question
>>
>>
>> On 6 Mar 2006, at 16:40, Iulian Alecu wrote:
>>
>>> Uli,Hai,John
>>>
>>> Thanks for your answers and sorry for multiple posting.
>>>
>>> Uli's explanation made me understand. The thing that bothered me was
>>> that I have an "or" who's acting as an "and" (well or as I just sad
>>> I'm missing a point).
>>>
>>> John in his answer proposing a model having intersections :
>>>
>>> Male partial Person
>>> Cat partial Animal
>>>
>>> CP1 = hasChild some Person and hasPet some Animal
>>> CP2 = hasChild some Male and hasPet some Cat
>>>
>>> CP3 = hasChild some Male and hasPet some owl:Thing
>>>
>>> (an more proper examples, thanks John), but actually he used  
>>> replaced
>>> by "and" the "or" in my example.
>>>
>>> Actually, this is the behavior that I expect from a and model not
>>> from a "or".
>>> Could anyone tell me why "or" and "and" looks like having the same
>>> result ?
>>>
>>
>> because they dont' matter in this case: what matters is that every  
>> cat
>> is a thing, but not every thing is a cat.....
>>
>> Cheers, Uli
>>
>>> Regards,
>>>
>>> iulian
>>>
>>>
>>> -----Message d'origine-----
>>> De : [hidden email]
>>> [mailto:[hidden email]] De la part de Uli
>>> Sattler Envoyé : lundi 6 mars 2006 16:49 À :
>>> [hidden email] Objet : [protege-owl] Re: reasonning
>>> question
>>>
>>>
>>> On 6 Mar 2006, at 14:52, Iulian Alecu wrote:
>>>
>>>>
>>>> Hi,
>>>>
>>>> I have the following situation for which I can't give my self an
>>>> explanation
>>>> :
>>>>
>>>> B partial A;
>>>> D partial C;
>>>>
>>>> Restrictions R1 and R2 ;
>>>>
>>>> CP1 complete (R1 some A) or (R2 some C);
>>>>
>>>> CP2 complete (R1 some B) or (R2 some D);
>>>>
>>>> At this point I agree that CP2 inferred_subclassof CP1;
>>>
>>> I agree as well
>>>
>>>>
>>>> Now, if I define
>>>> CP3 complete (R1 some B) or (R2 some owl:Thing);
>>>>
>>>
>>>>  I would expect to get after classification that both CP1 and CP2
>>>> are subclasses of CP3, bu actually I get just CP2 as subclass of
>>>> CP3.
>>>
>>> indeed, this is correct
>>>
>>>>  I feel that I'm missing an important point about reasonning in OWL
>>>> but I just can't figure out which.
>>>
>>> ok, so you say that
>>>
>>> - every instance of CP2 has an R1-successor in B or an R2- successor
>>> in D, and that
>>> - every instance of CP3 has an R1-successor in B or an R2-successor
>>> that can be anything.
>>>
>>> Hence CP2 is a subclass of CP3 (because being an instance of D
>>> clearly implies being an instance of anything/owl:Thing)
>>>
>>> CP3 is NOT a cubclass of CP2 because we can think of some x which is
>>> an instance of CP3 because it has an R2-successor y which is an
>>> instance of owl:thing, but not of D. Now, if x has no R1-
>>> successor at
>>> all (which would be fine for it being an instance of CP3), then  
>>> x is
>>> NOT an instance of CP2.
>>>
>>> Does this help? Cheers, Uli
>>>
>>>>
>>>> Thanks ,
>>>>
>>>> Iulian Alecu
>>>> -----------------------------------
>>>> Laboratoire SPIM
>>>> INSERM U729
>>>> -----------------------------------
>>>>
>>>> -------------------------------------------------------------------
>>>> -
>>>> -
>>>> -
>>>> ---
>>>> To unsubscribe go to http://protege.stanford.edu/community/
>>>> subscribe.html
>>>>
>>>
>>> --------------------------------------------------------------------
>>> -
>>> -
>>> ---
>>> To unsubscribe go to http://protege.stanford.edu/community/
>>> subscribe.html
>>>
>>> --------------------------------------------------------------------
>>> -
>>> -
>>> ---
>>> To unsubscribe go to http://protege.stanford.edu/community/
>>> subscribe.html
>>>
>>
>> ---------------------------------------------------------------------
>> -
>> ---
>> To unsubscribe go to http://protege.stanford.edu/community/
>> subscribe.html
>>
>> ---------------------------------------------------------------------
>> -
>> ---
>> To unsubscribe go to http://protege.stanford.edu/community/
>> subscribe.html
>>
>> ---------------------------------------------------------------------
>> -
>> ---
>> To unsubscribe go to http://protege.stanford.edu/community/
>> subscribe.html
>>
>> ---------------------------------------------------------------------
>> -
>> ---
>> To unsubscribe go to http://protege.stanford.edu/community/
>> subscribe.html
>>
>
> ----------------------------------------------------------------------
> ---
> To unsubscribe go to http://protege.stanford.edu/community/ 
> subscribe.html
>
> ----------------------------------------------------------------------
> ---
> To unsubscribe go to http://protege.stanford.edu/community/ 
> subscribe.html
>

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

12