HR: 0800h
AN: IN31B-1144    [Abstracts]
TI: The Semantic SPASE
AU: Hughes, S
EM: steve.hughes@jpl.nasa.gov
AF: Jet Propulsion Laboratory, MS 169-315 4800 Oak Grove Dr., Pasadena, CA 91109 United States
AU: Crichton, D
EM: dan.crichton@jpl.nasa.gov
AF: Jet Propulsion Laboratory, MS 169-315 4800 Oak Grove Dr., Pasadena, CA 91109 United States
AU: Thieman, J
EM: thieman@nssdc.gsfc.nasa.gov
AF: NASA/Goddard Space Flight Center, Code 690.1, Greenbelt, MD 20771 United States
AU: * Ramirez, P
EM: paul.ramirez@jpl.nasa.gov
AF: Jet Propulsion Laboratory, MS 169-315 4800 Oak Grove Dr., Pasadena, CA 91109 United States
AU: King, T
EM: tking@igpp.ucla.edu
AF: UCLA IGPP, BOX 951567, 5881 Slichter, Los Angeles, CA 90095 United States
AU: Weiss, M
EM: michele.weiss@jhuapl.edu
AF: The Johns Hopkins University Applied Physics Laboratory, 11100 Johns Hopkins Rd., Laurel, MD 20723-6099 United States
AB: The Semantic SPASE (Space Physics Archive Search and Extract) prototype demonstrates the use of semantic web technologies to capture, document, and manage the SPASE data model, support facet- and text-based search, and provide flexible and intuitive user interfaces. The SPASE data model, under development since late 2003 by a consortium of space physics domain experts, is intended to serve as the basis for interoperability between independent data systems. To develop the Semantic SPASE prototype, the data model was first analyzed to determine the inherit object classes and their attributes. These were entered into Stanford Medical Informatics' Protege ontology tool and annotated using definitions from the SPASE documentation. Further analysis of the data model resulted in the addition of class relationships. Finally attributes and relationships that support broad-scope interoperability were added from research associated with the Object-Oriented Data Technology task. To validate the ontology and produce a knowledge base, example data products were ingested. The capture of the data model as an ontology results in a more formal specification of the model. The Protege software is also a powerful management tool and supports plug-ins that produce several graphical notations as output. The stated purpose of the semantic web is to support machine understanding of web-based information. Protege provides an export capability to RDF/XML and RDFS/XML for this purpose. Several research efforts use RDF/XML knowledge bases to provide semantic search. MIT's Simile/Longwell project provides both facet- and text-based search using a suite of metadata browsers and the text-based search engine Lucene. Using the Protege generated RDF knowledge-base a semantic search application was easily built and deployed to run as a web application. Configuration files specify the object attributes and values to be designated as facets (i.e. search) constraints. Semantic web technologies provide the means to easily implement semantic search for science data archives using space science data models and metadata. It also suggests the means to support correlative search across science disciplines, missions, and instruments since multiple ontologies can be gathered together for processing . Finally broad scope interoperability can be envisioned where semantically aware software agents reason about and process distributed science data repositories.
DE: 0510 Agent-based models
DE: 0525 Data management
DE: 0599 General or miscellaneous
SC: Earth and Space Science Informatics [IN]
MN: Fall Meeting 2005