Design and Creation of Ontologies for Environmental Information Retrieval1

1. Copyright MCC and Telcordia Technologies, Inc., 1999. Domain Specific Ontologies are increaingly being used for information gathering and analysis over distributed and networked information sources. An example of such a system is the InfoSleuth system, where agents interpret user queries based on a domain specific ontology and cooperate with each other to retrieve the data corresponding to a query. We discuss the EDEN system for environmental information, where the ontology thus plays a crucial role. We propose an approach for designing an ontology for information retrieval based on: (a) the schemas of the databases; and (b) a collection of queries that are of interest to the users. In some cases, standardized vocabularies and data dictionaries may also aid the design and creation of the ontology. Our approach is presented in the context of the environmental information domain and minimizes the need for involvement of a domain expert for ontology construction. We discuss the EDEN ontology which was developed based on the environmental databases, CERCLIS 3, ITT, ERPIMS, HAZDAT, the data identified by the Basel Convention; and the set of user queries proposed by the user group. We discuss how an environmental data dictionary, the EDR repository and a standardized vocabulary such as the GEMET thesaurus can be used to improve the EDEN ontology.