Tutorial: Reasoning with 15926 Reference Data
Use this as the main page for planning the Semantic Technology tutorial for Semantic Days 2008.
- Introduction: Building ontology (no individuals). Emphasis on complexity.
- Reference data -- the RDL as a taxonomy
- With IDS methods, we can consolidate traditional (e.g., tabular) data in a standard RDL-respecting format
- We can extract subsets of a database for reasoning, using non-reasoning criteria for selection
- We can use sophisticated reasoning on such subsets.
Creating a reasonable representation of products in the oil and gas industry
A tutorial for Norwegian semantic days 2008
Session 1) Describing products as semantic models
1.1 Describe a simple everyday artifact (e.g., a cup)
Use different perspectives (geometry, materials, properties (thermal, strength), manufacturing, usage, maintenance (i.e., disposal) Include life-cycle cost attribute in all perspectives
1.2 Include also "real" examples from Drilling and Completion (Henning Jansen) and/or production and operation (e.g, pipeline example)
Session 2) Representing product models using reference data
Use ISO 15926 RDL tools (e.g., RD Browser, static html pages, etc.) to connect the product model to a standard representation Create additional relationships to transform the product model into a "product ontology"
Use above "real" examples to create lots of real-world individuals Connect individuals to geographic information Create mash-ups with WWW-applications (Google Earth, Del.icio.us, etc)
Generate XML file and use different XSLT scripts to display different representation formats (graphs, text, etc.)
Session 3) Reasoning about product properties and behavior
Use XSLT to transform the XML data (product ontology) into an OWL 1.1 file Display and maintain the product ontology in semantic editors (Protege, Swoop, etc.) Use SparQL to search extract relevant subset of information from large amounts of data about real-life individuals Use SWRL or other reasoning languages to reason about selected subset of the product ontology Reason (in Racer, Pellet, etc) to find "interesting" new "insights" about the product and/or its operation
In all of the above, show examples first, and then discuss what was done, how it whas done, what it implies, and how it can be used