Fwd: limiting number of subclasses: reg

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

Fwd: limiting number of subclasses: reg

sharmi m

Dear All,

Is it possible to define a class in protege as a class with exactly 2 number of subclasses?

If we define a third subclass, reasoner has to inform it as inconsistent.

Kindly suggest.


_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user
Reply | Threaded
Open this post in threaded view
|

Re: limiting number of subclasses: reg

Ralph Schäfermeier
Hi Sharmi,

You can make the superclass the disjoint union of its subclasses (see https://www.w3.org/TR/owl2-syntax/#Disjoint_Union_of_Class_Expressions).

However, adding a third subclass (that is disjoint with the two existing ones) would not render your ontology inconsistent. The new class would just be unsatisfiable. Only if you assert an instance of the new class, then the ontology becomes inconsistent.

Cheers,
Ralph


Am 08.06.2017 um 10:32 schrieb Sharmi M <[hidden email]>:


Dear All,

Is it possible to define a class in protege as a class with exactly 2 number of subclasses?

If we define a third subclass, reasoner has to inform it as inconsistent.

Kindly suggest.

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


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

signature.asc (465 bytes) Download Attachment
Reply | Threaded
Open this post in threaded view
|

Re: limiting number of subclasses: reg

Jim McCusker
That class would be satisfiable because of the non-unique naming assumption. It would simply be some other subset of the superclass that follows different criteria, but all of the members of that class must belong to one or the other of those disjoint union classes. However, if that class in question is also disjoint from its siblings, then it would be unsatisfiable, as you say.

If that doesn't work, I, at least, need a better sense of the use case for this.

Jim

On Thu, Jun 8, 2017 at 8:04 AM Ralph Schäfermeier <[hidden email]> wrote:
Hi Sharmi,

You can make the superclass the disjoint union of its subclasses (see https://www.w3.org/TR/owl2-syntax/#Disjoint_Union_of_Class_Expressions).

However, adding a third subclass (that is disjoint with the two existing ones) would not render your ontology inconsistent. The new class would just be unsatisfiable. Only if you assert an instance of the new class, then the ontology becomes inconsistent.

Cheers,
Ralph


Am 08.06.2017 um 10:32 schrieb Sharmi M <[hidden email]>:


Dear All,

Is it possible to define a class in protege as a class with exactly 2 number of subclasses?

If we define a third subclass, reasoner has to inform it as inconsistent.

Kindly suggest.

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

_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user
--
James P. McCusker III, Ph.D.

_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user
Reply | Threaded
Open this post in threaded view
|

Re: limiting number of subclasses: reg

Ralph Schäfermeier
I explicitly stated that the new subclass would have to be disjoint with the two existing ones. At least I assumed that that’s what the thread opener had in mind: a strict dichotomy of things where each individual is instance of either one class or the other (like, e.g,. even and odd numbers). You are right, we don’t know for sure, unless the enquirer tells us more about the use case.

Ralph


Am 08.06.2017 um 18:45 schrieb Jim McCusker <[hidden email]>:

That class would be satisfiable because of the non-unique naming assumption. It would simply be some other subset of the superclass that follows different criteria, but all of the members of that class must belong to one or the other of those disjoint union classes. However, if that class in question is also disjoint from its siblings, then it would be unsatisfiable, as you say.

If that doesn't work, I, at least, need a better sense of the use case for this.

Jim

On Thu, Jun 8, 2017 at 8:04 AM Ralph Schäfermeier <[hidden email]> wrote:
Hi Sharmi,

You can make the superclass the disjoint union of its subclasses (see https://www.w3.org/TR/owl2-syntax/#Disjoint_Union_of_Class_Expressions).

However, adding a third subclass (that is disjoint with the two existing ones) would not render your ontology inconsistent. The new class would just be unsatisfiable. Only if you assert an instance of the new class, then the ontology becomes inconsistent.

Cheers,
Ralph


Am 08.06.2017 um 10:32 schrieb Sharmi M <[hidden email]>:


Dear All,

Is it possible to define a class in protege as a class with exactly 2 number of subclasses?

If we define a third subclass, reasoner has to inform it as inconsistent.

Kindly suggest.

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

_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user
--
James P. McCusker III, Ph.D.
_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user


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

signature.asc (465 bytes) Download Attachment
Reply | Threaded
Open this post in threaded view
|

Re: limiting number of subclasses: reg

sharmi m
Thank You Dear Ralph & Jim. I think, i got your points..

 a super class is defined as disjoint union of 2 subclasses.

 While adding a new subclass, Inconsistency comes thro 2 ways.

1. by making the newly added subclass as disjoint class of existing 2  subclasses  or
2. by making the super class as disjoint class of existing 2 subclasses.


Thank you

On Fri, Jun 9, 2017 at 6:28 AM, Ralph Schäfermeier <[hidden email]> wrote:
I explicitly stated that the new subclass would have to be disjoint with the two existing ones. At least I assumed that that’s what the thread opener had in mind: a strict dichotomy of things where each individual is instance of either one class or the other (like, e.g,. even and odd numbers). You are right, we don’t know for sure, unless the enquirer tells us more about the use case.

Ralph


Am 08.06.2017 um 18:45 schrieb Jim McCusker <[hidden email]>:

That class would be satisfiable because of the non-unique naming assumption. It would simply be some other subset of the superclass that follows different criteria, but all of the members of that class must belong to one or the other of those disjoint union classes. However, if that class in question is also disjoint from its siblings, then it would be unsatisfiable, as you say.

If that doesn't work, I, at least, need a better sense of the use case for this.

Jim

On Thu, Jun 8, 2017 at 8:04 AM Ralph Schäfermeier <[hidden email]> wrote:
Hi Sharmi,

You can make the superclass the disjoint union of its subclasses (see https://www.w3.org/TR/owl2-syntax/#Disjoint_Union_of_Class_Expressions).

However, adding a third subclass (that is disjoint with the two existing ones) would not render your ontology inconsistent. The new class would just be unsatisfiable. Only if you assert an instance of the new class, then the ontology becomes inconsistent.

Cheers,
Ralph


Am 08.06.2017 um 10:32 schrieb Sharmi M <[hidden email]>:


Dear All,

Is it possible to define a class in protege as a class with exactly 2 number of subclasses?

If we define a third subclass, reasoner has to inform it as inconsistent.

Kindly suggest.

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

_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user
--
James P. McCusker III, Ph.D.
_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user


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



_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user
Reply | Threaded
Open this post in threaded view
|

Re: limiting number of subclasses: reg

Ralph Schäfermeier
Hi Sharmi,

Well, as I (and Jim) said, adding a third subclass and declaring it disjoint with the two existing ones does not necessarily make the ontology inconsistent, it only makes the new class unsatisfiable (i.e. a class that can never have an instance).  If the class has instances, then the ontology is inconsistent.

Not sure I understand your second point, but if you make your superclass disjoint with its two subclasses, then the two subclasses also become unsatisfiable (because a subclass can only have instances that are also instances of its superclass and at the same time, because of the disjointness, it cannot share any instance with its superclass).

As Jim also said, maybe it would be helpful if you gave a small example of what you want to achieve.

Cheers,
Ralph



Am 09.06.2017 um 19:06 schrieb Sharmi M <[hidden email]>:

Thank You Dear Ralph & Jim. I think, i got your points..

 a super class is defined as disjoint union of 2 subclasses.

 While adding a new subclass, Inconsistency comes thro 2 ways.

1. by making the newly added subclass as disjoint class of existing 2  subclasses  or
2. by making the super class as disjoint class of existing 2 subclasses.


Thank you

On Fri, Jun 9, 2017 at 6:28 AM, Ralph Schäfermeier <[hidden email]> wrote:
I explicitly stated that the new subclass would have to be disjoint with the two existing ones. At least I assumed that that’s what the thread opener had in mind: a strict dichotomy of things where each individual is instance of either one class or the other (like, e.g,. even and odd numbers). You are right, we don’t know for sure, unless the enquirer tells us more about the use case.

Ralph


Am 08.06.2017 um 18:45 schrieb Jim McCusker <[hidden email]>:

That class would be satisfiable because of the non-unique naming assumption. It would simply be some other subset of the superclass that follows different criteria, but all of the members of that class must belong to one or the other of those disjoint union classes. However, if that class in question is also disjoint from its siblings, then it would be unsatisfiable, as you say.

If that doesn't work, I, at least, need a better sense of the use case for this.

Jim

On Thu, Jun 8, 2017 at 8:04 AM Ralph Schäfermeier <[hidden email]> wrote:
Hi Sharmi,

You can make the superclass the disjoint union of its subclasses (see https://www.w3.org/TR/owl2-syntax/#Disjoint_Union_of_Class_Expressions).

However, adding a third subclass (that is disjoint with the two existing ones) would not render your ontology inconsistent. The new class would just be unsatisfiable. Only if you assert an instance of the new class, then the ontology becomes inconsistent.

Cheers,
Ralph


Am 08.06.2017 um 10:32 schrieb Sharmi M <[hidden email]>:


Dear All,

Is it possible to define a class in protege as a class with exactly 2 number of subclasses?

If we define a third subclass, reasoner has to inform it as inconsistent.

Kindly suggest.

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

_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user
--
James P. McCusker III, Ph.D.
_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user


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


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


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

signature.asc (465 bytes) Download Attachment
Reply | Threaded
Open this post in threaded view
|

Re: limiting number of subclasses: reg

sharmi m
Dear Ralph,

I realized the typo in my earlier statement  (point 2)

written as:   2. by making the super class as disjoint class of existing 2 subclasses.

To be corrected as :  2. by making the super class as disjoint class of  THE NEW SUBCLASS.

I am making sample models using protege to understand the basic concepts.
When I learnt about nominal class ( a class with single instance), In view of limiting  the instances,  I was thinking of the possible way to limit the number of subclasses.  That's  the background.

Thank you once again.

By the way,  I have already posted a question to our forum 

Are rules implemented thro' protege (windows->views->ontology views->rules) decidable?  Learned from literature that SWRL rules are undecidable? (which can be implemented thro SWRLTAB in protege).


- Sharmi






On Sat, Jun 10, 2017 at 4:58 AM, Ralph Schäfermeier <[hidden email]> wrote:
Hi Sharmi,

Well, as I (and Jim) said, adding a third subclass and declaring it disjoint with the two existing ones does not necessarily make the ontology inconsistent, it only makes the new class unsatisfiable (i.e. a class that can never have an instance).  If the class has instances, then the ontology is inconsistent.

Not sure I understand your second point, but if you make your superclass disjoint with its two subclasses, then the two subclasses also become unsatisfiable (because a subclass can only have instances that are also instances of its superclass and at the same time, because of the disjointness, it cannot share any instance with its superclass).

As Jim also said, maybe it would be helpful if you gave a small example of what you want to achieve.

Cheers,
Ralph



Am 09.06.2017 um 19:06 schrieb Sharmi M <[hidden email]>:

Thank You Dear Ralph & Jim. I think, i got your points..

 a super class is defined as disjoint union of 2 subclasses.

 While adding a new subclass, Inconsistency comes thro 2 ways.

1. by making the newly added subclass as disjoint class of existing 2  subclasses  or
2. by making the super class as disjoint class of existing 2 subclasses.


Thank you

On Fri, Jun 9, 2017 at 6:28 AM, Ralph Schäfermeier <[hidden email]> wrote:
I explicitly stated that the new subclass would have to be disjoint with the two existing ones. At least I assumed that that’s what the thread opener had in mind: a strict dichotomy of things where each individual is instance of either one class or the other (like, e.g,. even and odd numbers). You are right, we don’t know for sure, unless the enquirer tells us more about the use case.

Ralph


Am 08.06.2017 um 18:45 schrieb Jim McCusker <[hidden email]>:

That class would be satisfiable because of the non-unique naming assumption. It would simply be some other subset of the superclass that follows different criteria, but all of the members of that class must belong to one or the other of those disjoint union classes. However, if that class in question is also disjoint from its siblings, then it would be unsatisfiable, as you say.

If that doesn't work, I, at least, need a better sense of the use case for this.

Jim

On Thu, Jun 8, 2017 at 8:04 AM Ralph Schäfermeier <[hidden email]> wrote:
Hi Sharmi,

You can make the superclass the disjoint union of its subclasses (see https://www.w3.org/TR/owl2-syntax/#Disjoint_Union_of_Class_Expressions).

However, adding a third subclass (that is disjoint with the two existing ones) would not render your ontology inconsistent. The new class would just be unsatisfiable. Only if you assert an instance of the new class, then the ontology becomes inconsistent.

Cheers,
Ralph


Am 08.06.2017 um 10:32 schrieb Sharmi M <[hidden email]>:


Dear All,

Is it possible to define a class in protege as a class with exactly 2 number of subclasses?

If we define a third subclass, reasoner has to inform it as inconsistent.

Kindly suggest.

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

_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user
--
James P. McCusker III, Ph.D.
_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user


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


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


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



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