OWL and Existing Thesaurus Migration

1
OWL and Existing Thesaurus - Migration

• Marjorie M.K. Hlava
• Standards Implementer
• Access Innovations, Inc.
• Data Harmony Software

2
The Main Players

• Semantic Web standards
• XML, RDF, and OWL
• NISO, ISO, BSI
• Provide the framework
• Share across the web
• Platform independent
• Software independent

3
Thesaurus Standards

• Z39.19 Monolingual Thesauri
• ISO 2788 Monolingual
• ISO 5767 Multilingual
• A taxonomy is the hierarchical view of a
  thesaurus with information objects attached at
  the appropriate node

4
Data Base Standards

• Output
• MARC
• Format b
• RDF
• Markup
• XML
• SGML
• Metadata
• Dublin Core
• NFAIS /DOI Minimum data set
• Dialog default and additional fields

5
Fitting the Pieces Together

• XML - Rules, Syntax for Structured Documents
• RDF - Data Framework
• OWL concept records for the Web

6
eXtensible Markup Language

• XML syntax
• Markup
• Add content to the markup
• Attributes
• Elements
• Tag formats
• Well formed
• Transferable data

7
Resource Description Framework

• RDF
• standard method - simple descriptions
• RDF semantics - a clear set of rules for
  providing simple descriptive information
• RDF Schema a way for those descriptions to be
  combined into a single vocabulary.
• Self documenting XML

8
Whats an Ontology?

• subject - or domain - specific vocabularies.
• An ontology defines the terms used to describe
  and represent an area of knowledge.
• Ontologies include computer-usable definitions of
  basic concepts in the domain and the
  relationships among them.
• They encode knowledge in a domain and also
  knowledge that spans domains.

9
OWL

• Facilitates greater machine interpretability of
  Web content by providing additional vocabulary
  along with formal semantics.
• Tags in fielded format
• The fields add information about how to use that
  term
• Coordinates terms (pre-coordinate)

10
How OWL

• OWL uses
• URIs for naming
• RDF for description framework
• Capabilities to ontologies
• Ability to be distributed across many systems
• Scalability
• Compatibility with Web standards for
  accessibility and internationalization
• Openness
• Extensibility

11
OWL - Web Ontology Language

• Provides a language
• OWL Lite
• 10 properties
• 43 classes
• OWL DL and OWL Full
• 4 additional properties
• 11 more classes
• Defined for WWW architecture
• Semantic Web in particular.

12
Exactly what?

• OWL properties and classes
• relations between classes (e.g. disjointness),
• cardinality (e.g. "exactly one"),
• equality,
• richer typing of properties,
• characteristics of properties (e.g. symmetry),
• enumerated classes
• OWL Web Ontology Language Overview
• http//www.w3.org/TR/2004/REC-owl-features-2004021
  0/

13
So how are they different?

• Might not be
• Thesaurus Knowledge domain
• Different standards
• Different historical applications
• Ontologies in web
• Thesaurus in Data Bases
• OWL uses qualifiers to disambiguate terms

14
Tell me that in Thesaurus Terms

• Relations between classes (e.g.
  disjointness),Facets
• Cardinality (e.g. "exactly one"), Instance
• Equality, Use and USED FOR / Synonyms
• Richer typing of properties, More fields for user
  defining the terms
• Characteristics of properties (e.g. symmetry),
  Roles and treatment codes
• Enumerated classes - notation

15
Terms and concepts

• The stated differences
• Terms represent a concept
• Concepts are interrelated
• Clarity and maintenance challenges
• Ontologies seek to make it easier..

16
Migrating a thesaurus

• Digital form (not paper)
• Understand the construct of the thesaurus
• Get permission to use it
• Make it XML with RDF headers
• Keep the original thesaurus structure
• BT, RT PT notes etc
• Convert the labels to rdf format
• Rdfsscopenote
• Do not keep the node labels you dont need them

17
Migrating a thesaurus

• Try to use the WebOnt DTD for the XML
• http//www.w3c.org/2001/sw/WebOnt/web-ont-issues.h
  tmlI4.3-Structured-Datatypes
• Keep the original meanings

18
Semanitc conversion (?)

• Augment term to make it linkable and more concise
• Broader term owlTransitiveProperty
• Related term owlSymmetricProperty
• Preferred term prefLabel
• Non preferred term altLabels
• TopConcept is not top terms but special classes

19
In Ontologies

• Descriptors contain concepts
• Concepts are a set of terms
• One concept is a preferred concept of a
  descriptor
• Each descriptor can have qualifiers indicating
  aspects of a descriptor
• ABDOMEN -
• Anatomy
• Diseases
• Concepts in one descriptor are hierarchically
  related

20
NISO versus OWL

• Concepts versus terms
• Descriptors as a set of terms vs. a preferred
  term
• Qualifiers are not used to disambiguate terms
  use additional properties instead
• Relations as nodes
• Relations as arcs

21
Thesaurus fields

• Hierarchical
• BT / NT
• Associative
• Related Terms RT
• Equivalence
• Preferred terms Descriptors
• Use / Used For
• Notes
• Qualifiers and facets

22
How do I make my thesaurus into an Ontology?

• Map the field labels
• Add additional data
• Convert to XML
• Output the file in RDF instead of Native XML
• Then its an ontology

23
National Cancer Institute
24
NCI Ontology Record
25
NCI RDF DTD
26
Object Property DTD Sample

• Biochemical_Class_or_Structure" C74
  74 and_Drugs_Kind"/ cals_and_Drugs_Kind"/
  art_of_Chemical_or_Drug" C75
  75 and_Drugs_Kind"/ cals_and_Drugs_Kind"/
  Target_Organism" C73 73
  nd"/
  rdfID"rChemical_or_Drug_Has_Target_Anatomy"
  C71 71 rdfresource"Chemicals_and_Drugs_Kind"/
  
  rdfID"rChemical_or_Drug_FDA_Approved_for_Disease
  " C70 70 rdfresource"Chemicals_and_Drugs_Kind"/
  Kind"/

27
Semantic links

• se" C79 79 rdfresource"Organism_Kind"/ rdfresource"Findings_and_Disorders_Kind"/
  

28
Implied Attributes

• rdfID"rEO_Models_Human_Disease"
• C79
• 79
• Kind"/

29
Scope notes

• Concept scope note
• scopeNote is a rdfssubPropertyOf rdfscomment

30
Conclusion

• OWL is a standard to watch
• Implementation requires knowledge of UNIX / Linux
• Conversion from a current thesaurus is not
  difficult
• The base knowledge for applying requires new way
  of looking at concepts

31
Stay tuned for updates

• Marjorie M.K. Hlava
• President
• Access Innovations Inc
• Data Harmony Software
• 505-998-0800
• www.accessin.com
• www.dataharmony.com
• mhlava_at_accessinn.com