Research Plan: Metadata Knowledge Representation Using Topic Maps

1
Research Plan Metadata Knowledge Representation
Using Topic Maps

• Investigate current state of research in the
  following areas
• Topic Map standards - ISO/IEC 13250, XTM 1.0,
  TMQL
• Metadata Models (e.g. KDM, functional data
  models, etc.)
• Multidimensional Ontology representations
• Web standards (RDF, OIL, DAML, XML)
• Knowledge Management, Knowledge Learning
  Organization
• Resources
• Web Searches of standards organizations, academic
  and commercial sites (e.g. ISO, XTM, IEEE, ACM)
• Information exchange with subject matter experts
  (e.g. discussions, email exchange, visits of
  possible)
• Information exchange with student researchers who
  have similar interests
• Create a metadata model and develop a scaled-down
  prototype that demonstrates proposed concepts.

2
Issue to be Resolved
SUBJECT
Shipping Industry
South American World Events
ASSC C
Single Occurrences Linked Together
1
Topic A
Topic B
A
B
2
TEMPORAL (DATE/TIME)
3
OCC 1
OCC 2
Ontology and Topic overlap exists in several
dimensions
SPATIAL (REGION)
OCC 3
Topic Map
3
Research Scope Are the plans specific and
implementable

• Objective 1 Extend the Topic Map metadata model
  to include temporal, spatial and categorical
  information objects.
• Identify ways to interpret or extend the XTM 1.0
  standard
• Objective 2 Extend the Topic Map model to
  enable the linking of occurrences in an event
  chain for knowledge creation.
• Identify extensions to the XTM 1.0 standard
• Identify database schemas that work with XTM 1.0
  modifications to represent new knowledge gained
  via event chain analysis
• Objective 3 Investigate the spatial
  characteristics of occurrences and how the number
  of intermediate links affect the relevancy
  between the two.
• Create ontologies and measure spatial distances
  (I.e. number of links) of occurrences within the
  same and in different ontologies
• Develop a way to represent Topic Map links for
  relevancy analysis
• Determine ways in which visualization might help
  in the identification of same or similar
  occurrences