ONTOLOGY

1
ONTOLOGY (Tutorial I)
Arash Shaban-Nejad Comp 691B Foundation of
Semantic Web Winter 2004
2
OUTLINE

• Ontology Definition
• Ontologies Development Strategy
• Requirement Analysis
• Knowledge Acquisition
• Ontologies and knowledge base
• Evaluation
• References

3
Ontologies in Computer Science
Gruber An ontology is
Explicit specification of a Conceptualization
4
Ontologies Development Strategy
1. Specification 2. Knowledge acquisition 3.
Conceptualization 4. Integration 5.
Implementation 6. Evaluation 7. Documentation
5
Specification (Requirement Analysis)

• What is the goal of the ontology?(what is the
  usage?, users specifications )
• What is relevant to fulfill the goal (like
  entities,relationships, restrictions)?
• What need to be modeled? (E.g., key components
  of car, types of car)
• What granularity is useful? (What parts should
  be described, what is unnecessary ?)

6
Knowledge Acquisition

• Try to get the information based on the
  available documents in different data sources.

• Put the information in a hierarchy structure
  with respect to the ontology scope

• This step occurs in parallel with step (1)

7
Conceptualization
Concepts in the ontology should be close to
objects (physical or logical) and relationships
in your domain of interest.
8
Developing the ontology

• Domain terms will be identified as concepts and
  classes
• Arrange the classes in a taxonomic hierarchy to
  make the scope more efficient and abstract
• Define instances based on the classes
• Define slots and identify allowed values for
  them
• Fill in the values for slots for the instances

9
Integration
In order to obtain some uniformity across your
ontology with other ontologies, try to get
definitions from other ontologies
10
Implementation

• Ontology editor survey results
• Protégé 2000 is an ontology editor and a
  knowledge-base editor. It is also an open-source,
  Java tool that provides an extensible
  architecture for the creation of customized
  knowledge-based applications
• The latest version of Protégé 2000 supports OWL

11
Evaluation

• To check for Completeness , Consistence and
  avoiding from redundancy
•   Ontology nature is a iterative processing, So
  after evaluation we came back to previous phases
  and corrected mistakes.
• RACER is a description logic reasoning system
  with support for T-Box and A-Box

12
Description logic for testing ontologies
Primitives
Descriptions
Definitions
Reasoning
Validating
Thing
red partOf Car
13
Ontology and Knowledge Base
14
Review the steps
Ontology Purpose, Goals, Scope
Experts interview results, text analyses
Domain concepts, instances, relations and
properties
Integration with another ontology
Evaluation for Completeness Consistence redundancy
Ontology Implementation
Documentation
15
References

• University of buffalo http//ontology.buffalo.edu
  
• University of Manchester http//www.cs.man.ac.uk
  /horrocks/
• Deborah L. Mcguinness from Stanford university
  http//protege.stanford.edu/publications/ontology_
  development/ontology101-noy-mcguinness.html
• Developing Ontologies for medical domain by Dr.
  Rogers
• http//www.xmluk.org/public/magic-cicle-20
  02.htmmedical
• S. Staab and A. Maedche From Karlsruhe University
• http//events.aifb.uni-karlsruhe.de/ontol
  ogytutorial