LOMGen

1
LOMGen
Learning Object Metadata Generator
Features
Architecture

• Semi-automatic metadata generation,
• applied to Learning Object Metadata
• Shallow parsing of HTML objects to extract
• keywords/keyphrases
• Finds related terms and phrases, which are
• not explicit in the Learning Object
• LOMGen acts as an AgentMatcher
• component for indexing Learning Objects
• and for querying with keywords/keyphrases

Retrieved HTML file from LOR
Metadata Administrator
Fills in remaining tag values
Uses template, updates general identifier
Retrieve
Update
Extracted keyphrases, description and title

• Current work
• Extended parsing to derive weighted metadata
  queries
• Part-of-Speech Tagging
• WordNet interface for text analysis

Researchers Dr. Virendra C. Bhavsar Dr.
Harold Boley Anurag Singh
Faculty of Computer Science, University of New
Brunswick, P.O. Box 4400, 540 Windsor Street,
Gillin Hall E126, Fredericton, NB E3B 5A3 Phone
[phone] Fax [phone] Email
fcs_at_unb.ca
June 16, 2004