Extracting%20Semantic%20Relationships%20Between%20Wikipedia%20Articles

1
Extracting Semantic Relationships Between
Wikipedia Articles
Lowell Shayn Hawthorne Suzette Stoutenburg Superv
isor Jugal Kalita
University of Colorado at Colorado Springs (USA)
2
Acquiring Relationships Between Wikipedia Articles

• Wikipedia is arguably the largest source of
  collaboratively developed knowledge in the world
• But Wikipedia is largely unstructured and
  therefore unavailable for use by most software
  systems
• In the recent years, there has been increasing
  research in the use of Wikipedia as a broadly
  applicable lexical semantic resource
• Most approaches extract information from
  Wikipedia by harnessing implicit semantics in the
  syntactic structures 4, 10
• One approach has been proposed to explicitly
  express the relationships between links 16
  though this has not yet been implemented
• Other approaches use natural language processing
  techniques to extract knowledge from the
  structures of Wikipedia 5, 10, 14, 15 but none
  of these have focused on assigning meaning to the
  links between articles

3
Approach

• We extract the meaning of relationships between
  Wikipedia articles using natural language
  processing techniques
• We use regular expressions to detect linguistic
  patterns and infer relationships between each
  linked article pair
• Preliminary results are competitive for some
  relationships with high precision
• Use of regular expressions over parts of speech
  is feasible for knowledge extraction

4
Use patterns to acquire relationships
5
We All Want a Good Relationship

• Ontology alignment can support service
  composition, service mediation and enterprise
  integration
• Ontology alignment of key words and metadata
  about a users web searches could link buyers to
  desired products and fulfill sales online