How to specify individual name's uniqueness?

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

How to specify individual name's uniqueness?

Jason Wang
Hi,

How to specify individual name's uniqueness in Protege 4? I am thinking of how to detect conflicting assertions (e.g., contradictions resulting from cardinality as in the example I show below).

I found that when adding a new individual, I could set it "Different Individuals" in the Description part. This will in fact set all the related individuals to be "Different". But it did not change anything. Instead of showing contradictions, Plugin reasoners still infer implicitly the same results: conflicting assertions of object properties will lead to conclude two related individuals are same (explicitly it shows that two individuals have all the object properties of each other).

I tried the plugin reasoner Pellet and the pre-installed ones.

======

For example, I have a class A and a class B. Then I created one property "hasP" in "Object Properties" window, with domain A and range B. This property was set in the Restriction Type window with "Exactly (exact cardinality) = 1". Then there are a set of instances for A and B, such as {a1, a2, b1, b2, b3...}. However, when I add property assertions for "a1" and "b1":

1, a1 hasP b1
2, a2 hasP b2
3, a1 hasP b2

This will violate a1's hasP's cardinality of being 1 (because its "hasP" with both b1 and b2). I tried to use some reasoner plugins, it will infer:

a2 hasP b1 (i.e., a1 and a2 are same implicitly)

_______________________________________________
p4-feedback mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/p4-feedback
Reply | Threaded
Open this post in threaded view
|

Re: How to specify individual name's uniqueness?

Lorenz Buehmann
Hi,

this behaviour of the reasoners is correct as there is theoretically no UNA(Unique Name Assumption) in OWL. But if there is explicitely assigned that b1 and b2 are different(owl:differentFrom), then it should result in an inconsistent ontology.

Regards,
Lorenz
On 01/03/2013 07:54 PM, Jason Wang wrote:
Hi,

How to specify individual name's uniqueness in Protege 4? I am thinking of how to detect conflicting assertions (e.g., contradictions resulting from cardinality as in the example I show below).

I found that when adding a new individual, I could set it "Different Individuals" in the Description part. This will in fact set all the related individuals to be "Different". But it did not change anything. Instead of showing contradictions, Plugin reasoners still infer implicitly the same results: conflicting assertions of object properties will lead to conclude two related individuals are same (explicitly it shows that two individuals have all the object properties of each other).

I tried the plugin reasoner Pellet and the pre-installed ones.

======

For example, I have a class A and a class B. Then I created one property "hasP" in "Object Properties" window, with domain A and range B. This property was set in the Restriction Type window with "Exactly (exact cardinality) = 1". Then there are a set of instances for A and B, such as {a1, a2, b1, b2, b3...}. However, when I add property assertions for "a1" and "b1":

1, a1 hasP b1
2, a2 hasP b2
3, a1 hasP b2

This will violate a1's hasP's cardinality of being 1 (because its "hasP" with both b1 and b2). I tried to use some reasoner plugins, it will infer:

a2 hasP b1 (i.e., a1 and a2 are same implicitly)


_______________________________________________
p4-feedback mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/p4-feedback


_______________________________________________
p4-feedback mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/p4-feedback
Reply | Threaded
Open this post in threaded view
|

Re: How to specify individual name's uniqueness?

Jason Wang
Lorenz, thank you so much!

Yes, I should specify that b1 and b2 are also different. Now the protege can detect inconsistency!!!


On Thu, Jan 3, 2013 at 1:50 PM, Lorenz B├╝hmann <[hidden email]> wrote:
Hi,

this behaviour of the reasoners is correct as there is theoretically no UNA(Unique Name Assumption) in OWL. But if there is explicitely assigned that b1 and b2 are different(owl:differentFrom), then it should result in an inconsistent ontology.

Regards,
Lorenz

On 01/03/2013 07:54 PM, Jason Wang wrote:
Hi,

How to specify individual name's uniqueness in Protege 4? I am thinking of how to detect conflicting assertions (e.g., contradictions resulting from cardinality as in the example I show below).

I found that when adding a new individual, I could set it "Different Individuals" in the Description part. This will in fact set all the related individuals to be "Different". But it did not change anything. Instead of showing contradictions, Plugin reasoners still infer implicitly the same results: conflicting assertions of object properties will lead to conclude two related individuals are same (explicitly it shows that two individuals have all the object properties of each other).

I tried the plugin reasoner Pellet and the pre-installed ones.

======

For example, I have a class A and a class B. Then I created one property "hasP" in "Object Properties" window, with domain A and range B. This property was set in the Restriction Type window with "Exactly (exact cardinality) = 1". Then there are a set of instances for A and B, such as {a1, a2, b1, b2, b3...}. However, when I add property assertions for "a1" and "b1":

1, a1 hasP b1
2, a2 hasP b2
3, a1 hasP b2

This will violate a1's hasP's cardinality of being 1 (because its "hasP" with both b1 and b2). I tried to use some reasoner plugins, it will infer:

a2 hasP b1 (i.e., a1 and a2 are same implicitly)


_______________________________________________
p4-feedback mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/p4-feedback


_______________________________________________
p4-feedback mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/p4-feedback



_______________________________________________
p4-feedback mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/p4-feedback