An Ontology Design Pattern for
Representing Relevance in OWL
Juan Gómez-Romero,
Fernando Bobillo, Miguel Delgado
University of Granada
Department of Computer Science and A.I.
ISWC2007
outline
•
•
•
•
•
Information overload
Ontology design pattern
Reasoning
Related work
Future work
2
information overload
• In general: People get more
information than understandable
• In Information Systems: Users get
overwhelmed by information
provided by the system
• In Mobile Systems: Easier to be
“overloaded” with information
3
example
• A doctor is attending to a patient outside the
hospital (e.g.: emergency units, primary
healthcare in remote areas)
• The doctor uses a portable device to connect
to the Hospital Information System (HIS)
• He gets a bunch of Electronic Health Records
(EHRs) about the patient in order to suggest a
treatment
• He filters the results manually and grasps
interesting information (if he has enough time
and knowledge)
• Overload!
4
relevance of information
• Solution: To provide only “relevant”
information
• But… what is relevant?
• It depends on:
–
–
–
–
User preferences
User environment
User previous actions
…
• Context (in a wide sense!)
– “The interrelated conditions in which something
exists or occurs” (Merryam-Webster)
5
example
• A doctor is attending to a patient outside
the hospital (e.g.: emergency units,
primary healthcare in remote areas)
• He is interested only in some pieces of
advice related with the patient situation
• For instance,
– If the patient is “slightly unconscious” and
has an “hemorrhagic laceration”…
– information about if he has been diagnosed
of “bad reactions to procaine” should be
taken into account
6
representation of relevance
• We need to represent “relevance”
(which depend on context) in
ontologies
• “Desirable”: Standard OWL
language and tools must be used
• Identify common problems and
provide some suggestions to solve
them → An ontology design pattern
7
CDR pattern (i)
• Context-Domain Relevance (CDR,
pronounced “cider”) pattern
• A pattern to build an OWL ontology which
represents relevance relations
• Context ontology: vocabulary to describe
scenarios/circumstances/etc.
• Domain ontology: knowledge about the
problem to be solved
• Profile ontology: Links among complex
contexts and domains
8
CDR pattern (ii)
Hospital Information System
HIS Abstract Model
(A ó B ) ò E  D 1
Emergency Situation
Description Model
P1,1   R 1.C 1 ó  R 2 .D 1
C 1  (F ó G )
Context-Domain Relevance Model
9
inference
•
Restricted domain I of a scenario S
– To get the domain information which is relevant in a given context
•
Algorithm
–
–
–
–
Get the complex contexts Cn more general than S
Get the profiles Pn,m involving Cn
Get the complex domains Dm involved in Pn,m
Get the concepts I of the domain more specific than Dm
Domain Ontology
Context Ontology
I
Dm
I
Pn,m
Cn
S
Context-Domain Relevance Model
10
CDR plug-in for Protégé
11
features
• Reusability (pattern and ontologies)
• Standardization (basic requirement)
• Formalization (formal semantics of the
relevance ontology and the inference)
• Modularization (promotes ontology
modularization)
• Expressivity (enough to represent relevance)
• Complexity (bounded by context and domain
expressions)
12
related work
•
Ontology design patterns
– Svátek, V.(2004). Design Patterns for Semantic Web Ontologies:
Motivation and Dicussion
– Gangemi , A. (2005). Ontology Design Patterns for Semantic Web Content
•
Context/environment representation in Pervasive Computing
– Chen, H., Finin, T., Joshi, A. (2005). The SOUPA Ontology for Pervasive
Computing
– Gu, T., Pung, H., Zhang, D. (2005). A service-oriented middleware for
building context-aware services
– Lassila, O. , Khushraj, D. (2005). Contextualizing applications via semantic
middleware
•
Contextualization of ontologies
– Guha, R., McCool, R., Fikes, R. (2004). Contexts for the Semantic Web
– Bouquet, P., Giunchiglia, F., van Harmelen, F., Serafini, L.,
Stuckenschmidt, H. (2004). Contextualizing ontologies
– Stuckenschmidt, H. (2006). Toward Multi-viewpoint Reasoning with OWL
Ontologies
13
future work
• Promote pattern (and good practices)
specifications for ontology development
• Study and solve issues with
owl:import
• Enhancing and supporting Protégé
plug-in
• Fuzzy extension of the relevance
ontology for representing partial context
inclusion and ranking of relevance
relations
14
end
questions? comments?
thank you!
감사합니다
15
Descargar

www.giaa.inf.uc3m.es