Submissions:NegativePropertyAssertions
From Odp
Current revision (08:29, 31 May 2010) (view source) m (Text replace - 'WOP2009:Main' to 'WOP:2009') |
|||
(6 intermediate revisions not shown.) | |||
Line 1: | Line 1: | ||
{{Logical_OP_Proposal_toolbar}} | {{Logical_OP_Proposal_toolbar}} | ||
- | {{ | + | {{Graphical representation header}} |
+ | {{Graphical representation | ||
+ | |ImageName=Npa-diagram.png | ||
+ | }} | ||
{{Logical OP General Template | {{Logical OP General Template | ||
|Name=NegativePropertyAssertions | |Name=NegativePropertyAssertions | ||
Line 15: | Line 18: | ||
[2] Motik, B., Patel-Schneider, P.F., Parsia, B.: OWL 2 Structural Specification and Functional-Style Syntax. W3C Candidate Recommendation 11 June 2009, 2009. | [2] Motik, B., Patel-Schneider, P.F., Parsia, B.: OWL 2 Structural Specification and Functional-Style Syntax. W3C Candidate Recommendation 11 June 2009, 2009. | ||
|Aim=Expressing NPAs in ontologies prior to OWL 2 as well as given an transformation rule when using OWL 2. | |Aim=Expressing NPAs in ontologies prior to OWL 2 as well as given an transformation rule when using OWL 2. | ||
- | |Solution=NegativeObjectPropertyAssertion( | + | |Solution=NegativeObjectPropertyAssertion(prop i1 i2) is equivalent to (using OWL 2 Abstract Syntax): |
- | SubClassOf(ObjectOneOf(i1), ObjectComplementOf(ObjectSomeValuesFrom(prop, ObjectOneOf(i2))))) | + | SubClassOf(ObjectOneOf(i1), ObjectComplementOf(ObjectSomeValuesFrom(prop, ObjectOneOf(i2)))) |
+ | |||
+ | |||
+ | Let ''C'' and ''D'' be concepts. Then ''C'' and ''D'' are disjoint if, and only if, ''C'' is subsumed by the complement of ''D'', i.e., '( | ||
+ | SubClassOf( C ObjectComplementOf(D) ). | ||
+ | |||
+ | The equivalence is correct because of the duality of disjointness, equivalence, and unsatisfiability: ''C'' is subsumed by ''D'' if, and only if, | ||
+ | ObjectIntersectionOf( C ObjectComplementOf(D) ) is unsatisfiable, and the intersection of ''C'' and ''D'' is unsatisfiable if, and only if, ''C' and ''D'' are disjoint. | ||
+ | |||
+ | One also reminds that the extension of the concept | ||
+ | ObjectSomeValuesFrom(prop C) is the set of individuals ''i'' which are connected to an individual ''j'' that is in the extension of the concept ''C'', by the property ''prop''. | ||
+ | |||
+ | |||
+ | Let ''NegativePropertyAssertion(p a b)'' be a negative property assertion axiom, i.e., the individual ''a'' is not related to ''b'' by the property ''p''. Then the extension of | ||
+ | ''ObjectSomeValuesFrom( p ObjectOneOf(b) )'' which contain all individuals that are connected to ''b'' by ''p'' must not contain ''a''. This is true, if, and only if ''ObjectOneOf(a)'' is disjoint to ObjectSomeValuesFrom( p ObjectOneOf(b) )'' | ||
|Elements=Individiual i1 | |Elements=Individiual i1 | ||
Line 24: | Line 41: | ||
ObjectProperty prop | ObjectProperty prop | ||
}} | }} | ||
- | {{Logical OP Example Template}} | + | {{Logical OP Example Template |
+ | |ProblemExample=Consider a social network containing facts about people and their relationships. | ||
+ | Let ''Adam'' and ''Eve'' be two persons and like a property (''A likes B''). Furthermore we know that ''Adam'' does not like ''Eve'' but we have no dislike relationship. Moreover, our language (such as OWL 1) does not have any NPA axiom constructor. | ||
+ | The sample ontology is interpreted with respect to the open-world semantics, | ||
+ | i. e. , one can not infer the dislike merely from the lack of a property assertion axiom ''ObjectPropertyAssertion(like Adam Eve)''. Then this fact can be expressed with the following axiom (we will also use the OWL 2 Abstract Syntax here): | ||
+ | |||
+ | ''SubClassOf (Adam ComplementOf(ObjectSomeValuesFrom( likes ObjectOneOf(Eve))))'' | ||
+ | }} | ||
{{Logical OP Reference Template}} | {{Logical OP Reference Template}} | ||
+ | {{Additional information header}} | ||
[[Category:Review assigned]] | [[Category:Review assigned]] | ||
{{Scenarios about me}} | {{Scenarios about me}} | ||
{{Reviews about me}} | {{Reviews about me}} | ||
+ | {{Modeling issues about me}} | ||
+ | {{My references}} | ||
{{Submission to event | {{Submission to event | ||
- | |Event= | + | |Event=WOP:2009 |
}} | }} |
Current revision
If you are a member of quality committee please visit the
If you are author of this proposal or you want to contribute to this pattern's review, you can: specify if this revision takes in account any of the review(s) In general, it could be useful to visit the evaluation section to have information about the evaluation process of this proposal Current revision ID: 9705 |
Graphical representation
Diagram
General information
Name | NegativePropertyAssertions |
---|---|
Also known as | |
Author(s) | |
SubmittedBy | OlafNoppens |
Description
Motivation | The motivation of this pattern is to model negative property assertions (NPAs) in ontology languages such as OWL 1 [1] that do not provide a special constructor for expressing it. It is worth mentioning that not all knowledge base systems can be migrated to OWL 2 [2] for several reasons. On the other hand, NPAs modeled according to this pattern can be migrated to OWL 2 using the newly introduced constructor.
A negative property assertion as defined in the upcoming OWL 2 states that a given individual i is never connected to a given individual j by a given property expression P. In other words, asserting that i is connected to j by P results in an inconsistent ontology. In this sense this assertion can be considered as a constraint that should not be violated. In contrast, considering an ontology where it cannot be inferred that i is connected to j by P does not necessarily mean that there cannot be such a connection - in fact, it is merely not modeled.
[2] Motik, B., Patel-Schneider, P.F., Parsia, B.: OWL 2 Structural Specification and Functional-Style Syntax. W3C Candidate Recommendation 11 June 2009, 2009. |
---|---|
Aim | Expressing NPAs in ontologies prior to OWL 2 as well as given an transformation rule when using OWL 2. |
Solution description | NegativeObjectPropertyAssertion(prop i1 i2) is equivalent to (using OWL 2 Abstract Syntax):
SubClassOf(ObjectOneOf(i1), ObjectComplementOf(ObjectSomeValuesFrom(prop, ObjectOneOf(i2))))
The equivalence is correct because of the duality of disjointness, equivalence, and unsatisfiability: C is subsumed by D if, and only if, ObjectIntersectionOf( C ObjectComplementOf(D) ) is unsatisfiable, and the intersection of C and D is unsatisfiable if, and only if, C' and D are disjoint. One also reminds that the extension of the concept ObjectSomeValuesFrom(prop C) is the set of individuals i which are connected to an individual j that is in the extension of the concept C, by the property prop.
|
Elements | Individiual i1
Individual i2 ObjectProperty prop |
Implementation | |
Reusable component | |
Component type |
Example
Problem example | Consider a social network containing facts about people and their relationships.
Let Adam and Eve be two persons and like a property (A likes B). Furthermore we know that Adam does not like Eve but we have no dislike relationship. Moreover, our language (such as OWL 1) does not have any NPA axiom constructor. The sample ontology is interpreted with respect to the open-world semantics, i. e. , one can not infer the dislike merely from the lack of a property assertion axiom ObjectPropertyAssertion(like Adam Eve). Then this fact can be expressed with the following axiom (we will also use the OWL 2 Abstract Syntax here): SubClassOf (Adam ComplementOf(ObjectSomeValuesFrom( likes ObjectOneOf(Eve)))) |
---|---|
Pattern solution example | |
Consequences |
Pattern reference
Origin | |
---|---|
Known use | |
Reference | |
Related ODP | |
Used in combination with | |
Test |
Additional information
Scenarios
No scenario is added to this Content OP.
Reviews
Review article | Posted on | About revision (current is 9705) |
---|---|---|
ValentinaPresutti about NegativePropertyAssertions | 24550849 September 2009 | 57695,769 |
StefanoDavid about NegativePropertyAssertions | 245508510 September 2009 | 57785,778 |
RinkeHoekstra about NegativePropertyAssertions | 245512924 October 2009 | 59035,903 |
This revision (revision ID 9705) takes in account the reviews: none
Other info at evaluation tab
Modeling issues
There is no Modeling issue related to this proposal.
References
Submission to event |
---|